JPH07110811A - Character processor and its method - Google Patents

Abstract

PURPOSE:To provide a character processor and its method having excellent operability for the correcting processing of a misinput generated at the time of inputting a read character input. CONSTITUTION:The character processor is provided with a reading input part 1 for inputting a reading, an input reading string temperary storage part 2 for temporarily storing an inputted reading string, an extended candidate generating part 4 for generating an extended candidate by interpreting a partial reading string in the reading string stored in the storage part 2 by another reading, an extended candidate temporary storage part 5 for storing the extended candidate generated by the generating part 4, a display attribute updating part 8 for changing the display attributes of the extended candidate reading string stored in the storage part 5 only by a part different from the reading string stored in the storage part 2, and a candidate display part 14 for identifying and displaying the reading string of the extended candidate stored in the storage part 5 in accordance with the changed display attribute.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character processing device and a character processing method which are used in a word processing system such as a word processor or a personal computer and which is controlled by a digital computer and have a function of correcting an erroneous input of a Japanese input document.

[0002]

2. Description of the Related Art In recent years, a kana-kanji conversion method has become the mainstream as a Japanese input method. This method uses the reading of each character of the document as input data to the document processing apparatus to specify the notation of the sentence to be created. Then, when creating a document in which kana-kanji conversion processing uses a phonetic character string as input data, there are many erroneous instructions due to typos, incorrect kana usage, shift omission, etc., when inputting and instructing this reading. .

Therefore, conventionally, a kana-kanji conversion method has been devised for obtaining a desired kanji-kana mixed string even with such an incorrect reading input. FIG. 19 is a block diagram showing the configuration of a conventional character processing device having such an erroneous input correction function. In FIG. 2, reference numeral 1 is a reading input unit, which generates a code corresponding to a character keyed on the keyboard device.

Reference numeral 2 denotes an input reading string temporary storage unit which sequentially stores the character codes input from the reading input unit 1 and temporarily stores them as a series of character reading strings. Reference numeral 3 is a replacement table for storing a set of pseudonyms used in the semiconductor storage device, which are prone to errors. For example, "ji" and "ji", "zu" and "zu",
A group such as "to" and "to" is stored.

Reference numeral 4 is a replacement string generation unit, which searches the replacement table 3 for the readings stored in the input reading string temporary storage unit 2. Then, if there is a match, the retrieved reading in the input reading string is replaced with the corresponding reading.
A read sequence temporary storage unit 5 includes a read sequence generation unit 4
The temporary replacement result is temporarily stored as a replacement string.

A display device 6 is provided on the liquid crystal display device.
Display the character corresponding to the specified code. Reference numeral 7 denotes an attention candidate temporary storage unit that temporarily stores the sequence number of the conversion candidate currently being noticed among the plurality of conversion candidates. Reference numeral 8 denotes a conversion candidate generation unit, and when the "conversion key" on the keyboard is pressed, the input segment and the transposition column in the input reading sequence temporary storage unit 2 and the substitution sequence temporary storage unit 5 are separated into each phrase and Estimate conversion candidates in bunsetsu and generate candidates for kanji and kana mixed sequences with high probability.

A conversion candidate temporary storage unit 9 temporarily stores the conversion candidates generated by the conversion candidate generation unit 8. For example,
When a replacement sequence of "Koori" is generated for the input reading of "Kouri", "Retail (Kouri),
"Axiom (Kouri), high profit (Kouri), utilitarianism (Kouri), ice (Koori), mulberry (Koori), freeze (Koori)" is stored in association with each conversion candidate.

A candidate display unit 10 outputs the conversion candidates in the conversion candidate temporary storage unit 9 to the display device 6 and displays them on the liquid crystal screen. At this time, the attention candidate temporary storage unit 7 is referred to and the currently noticed transformation candidate among the plurality of transformation candidates is highlighted. If the contents of the attention candidate temporary storage unit 7 change, the notation of the conversion candidate corresponding to the conversion candidate number is displayed in reverse video.

Reference numeral 11 denotes a candidate selection unit, which is a candidate candidate temporary storage unit 7 when the "conversion key" on the keyboard device is pressed.
The above noted sequence number is sequentially incremented. If the current number is the same as the number of homophone candidates (maximum sequence number), then set to "1". Further, when there is a candidate instruction by keystroke of the "number key" or a target candidate instruction by keystroke of the "confirmation key" and "reading key", the corresponding candidate is determined to be the desired one.

Next, the operation of the above-mentioned conventional character processing device will be described. First, the reading input from the reading input unit 1 is temporarily stored in the input reading string temporary storage unit 2.
When the user presses the "conversion key", the reading sequence generating unit 4 refers to the reading table 3 to generate a reading sequence in which the partial reading sequence in the input reading sequence temporary storage unit 2 is replaced, and the reading sequence is generated. It is stored in the temporary storage unit 5.

After that, the conversion candidate generation unit 8 generates conversion candidates for the input reading sequence in the input reading sequence temporary storage unit 2 and the reading sequence in the reading sequence temporary storage unit 5, and stores them in the conversion candidate temporary storage unit 9. Store. Further, the leading candidate number “1” of the conversion candidates is written in the attention candidate temporary storage unit.
The candidate display unit 10 displays a list of only the candidates of the candidate number specified in the attention candidate temporary storage unit 7 on the display device 6 with an inversion attribute.

When the user presses the "conversion key", the candidate selection section 11 sequentially increments the numbers in the attention candidate temporary storage section 7, and accordingly the conversion candidate displayed on the display device 6 is highlighted. Change. As described above, for example, with respect to the input reading string “Kouri”,
If "kou" is read as "koo,"
The display shown in 0 is made. Therefore, even if you input "ice" (koori) by mistake and enter "kouri" in kana, the necessary candidates for conversion will be displayed, so you can easily select it by selecting the conversion candidate. You can enter the characters.

[0013]

However, the conventional character input device has the following inconveniences from the viewpoint of simplifying erroneous input correction during the input operation by the operator. (1) In the character processing device shown in FIG. 19, if the partial reading in the input reading string matches the reading target reading string in the reading table 3, a plurality of candidates may be detected. For example, in the case that the Japanese sound and the dakuon are read as the direct sound and the pure sound respectively, specifically, when you input "Kou try", "ko" becomes "go", "shi" becomes "ji", " As "yo" is converted into "yo", as shown in Fig. 21, it is for teachers, for negotiations, for construction, for improvement, for factories, and for factories. ), Joint venture (goshou), auspicious trade (gosho), and obstinacy (gojo) are presented as candidates. In this case, even if not only the kana-kanji character string obtained as the conversion result but also the reading string interpreted and read is also displayed, the display of the reading string is similar and it is difficult to distinguish them. Therefore, there is a high possibility that the operator will make a selection error when selecting the desired word notation. (2) In the conventional character processing device, when the user who is accustomed to the kana-kanji conversion operation selects the candidate even if the first candidate is the correction candidate, without being aware of the correction. There were many Therefore, a similar candidate that is not originally intended may be selected, and the correction candidate is displayed, which may cause a side effect of the correction that an inappropriate candidate is selected. (3) Generally, there are individual differences in the method of input operation, and those who use the sequential deterministic method that confirms everything at the time of conversion, and those who select the word notation and confirm the term after completing the input for the time being Some people use confirmation methods. However, in the conventional character processing device, once a candidate was selected and confirmed as a text, it was not possible to confirm which portion the conversion candidate for erroneous input correction was selected after the confirmation process. For this reason, when a person accustomed to the method of the post-confirmation operation performs the “confirmation of the correction result”, he / she has to perform the difficult operation of estimating the reading from the text and confirming the correctness of the correction. . (4) Further, in the conventional character processing device, the input reading string is difficult to exist as a reading of a Japanese word in the middle of the input, such as "Fukazoku ..." or "Arkorizumu". Even if it was input, it was left as it was until the operator instructed to start "Conversion". For this reason, the operator originally had to correct the input readings at the same time "when selecting a conversion candidate", which requires complicated operations such as punctuation and homophone selection. This increases the burden on the operator. It was

As described above, in the conventional character processing apparatus, "the visibility of the candidate is poor", "the confirmation of the correction fact is difficult", "the confirmation after confirmation is difficult", "the correction timing is late", etc.
There were various problems in order to improve the operability of the correction function for erroneously input characters. Therefore, the present invention has been made to solve the above problems, and provides a character processing device and a character processing method capable of improving the operability during correction processing of erroneous input that occurs when a reading character string is input. The purpose is to do.

[0015]

According to a first aspect of the present invention, there is provided a character processing device which comprises a phonetic input means for inputting a phonetic character string in which each character of a sentence to be created is represented by a phonetic character, and the input phonetic character. When the character string corresponds to a predetermined erroneous input mode, the extended reading character string is generated by replacing the corresponding portion in the reading character string with another reading character according to a predetermined rule, and the extended reading character string is generated. To the notation character string to generate an expansion candidate, and an extended candidate character string generated by the expanded candidate generation means and replaced with the reading character string replaced by the expanded candidate generation means. Attribute updating means for giving a display attribute different from that of the reading character string, and a candidate displaying means for displaying the extended reading character string according to the display attribute given by the display attribute updating means. Having.

In the character processing device according to the second aspect of the present invention, the reading input means for inputting the reading character string in which each character of the sentence to be created is represented by phonetic characters, and the input reading character string are predetermined. When it corresponds to the erroneous input mode, the extended reading character string is generated by replacing the corresponding portion in the reading character string with another reading character according to a predetermined rule, and the input reading character string and the extended reading character Expansion candidate generation means for generating expansion candidates by converting a string into a notation string, and display order setting means for setting the display order of the expansion candidates generated by the expansion candidate generation means according to a predetermined rule are set. And a candidate display unit for displaying the extended candidate and the extended reading character string in accordance with the display order.

The character processing apparatus according to the invention of claim 3 is different from the invention of claim 2 in that the extended reading character string of the extension candidate set to the first priority output order by the display order setting means is input. Among the reading character strings, the maximum likelihood candidate determining means for determining whether or not the character string corresponding to the extended reading character string is equal, and the maximum likelihood candidate determining means determines that the reading character string and the extended reading character string are different. When the determination is made, it has a warning output means for outputting a warning in a predetermined format.

In the character processing device according to the invention of claim 4, the phonetic input means for inputting a phonetic character string in which each character of the sentence to be created is represented by phonetic characters, and the phonetic character is input from the phonetic input means. Each time, the reading evaluation means for determining whether or not the input reading character corresponds to a character of a predetermined erroneous input form, and the reading character determined when the reading evaluation means determines the erroneous input character And a reading character correcting means for replacing the reading character with a preset alternative reading character.

In the character processing device according to the invention of claim 5, the reading input means for inputting the reading character string in which each character of the sentence to be created is represented by phonetic characters, and the input reading character string are predetermined. When it corresponds to the erroneous input mode, the extended reading character string is generated by replacing the corresponding portion in the reading character string with another reading character according to a predetermined rule, and the extended reading character string is converted into the written character string. The expansion candidate generating means for generating the expansion candidate by doing, the candidate display means for displaying the expansion candidate and the extended reading character string generated by the expansion candidate generating means, and the operator from the expansion candidates displayed by the candidate display means A candidate selection unit for selecting a desired extension candidate, the extension candidate, and extension history information indicating that the extension candidate is generated based on the extension reading character string are stored in association with each other. Based on the text storage means, the correction location search means for searching the location information in the converted sentence of the extension candidate associated with the extension history information from the text storage means, and the location information of the extension candidate searched by the correction location search means. And a search result display means for displaying the character displayed at the position indicated by the position information in a form that can be identified.

In the character processing method according to the invention of claim 6, a phonetic input step of inputting a phonetic character string in which each character of a sentence to be created is represented by phonetic characters, and the input phonetic character string is predetermined. When it corresponds to the erroneous input mode, the extended reading character string is generated by replacing the corresponding portion in the reading character string with another reading character according to a predetermined rule, and the extended reading character string is converted into the written character string. The extended candidate generation step for generating an extended candidate by doing the above, and the reading character string portion replaced in the extended candidate generation step among the extended reading character string generated in the extended candidate generation step, and a portion of another reading character string. Display attribute update step for assigning a display attribute different from the above, and candidates for displaying the extended reading character string according to the display attribute assigned in the display attribute update step And a step shows.

In the character processing method according to the invention of claim 7, a reading input step for inputting a reading character string in which each character of a sentence to be created is represented by phonetic characters, and the input reading character string is predetermined. When it corresponds to the erroneous input mode, the extended reading character string is generated by replacing the corresponding portion in the reading character string with another reading character according to a predetermined rule, and the input reading character string and the extended reading character An expansion candidate generation step of generating expansion candidates by converting a column into a notation character string, and a display order setting step of setting the display order of the expansion candidates generated in the expansion candidate generation step according to a predetermined rule are set. And a candidate display step of displaying the extended reading character string in accordance with the display order.

The character processing method according to the invention of claim 8 is different from the invention of claim 7, in that the extended reading character string of the extension candidate set to the first priority output order by the display order setting step is input. Among the read character strings, the maximum likelihood candidate determination step of determining whether or not the character string corresponding to the extended read character string is equal, and the read character string and the extended read character string are different in the maximum likelihood candidate determination step. And a warning output step for outputting a warning in a predetermined format when the determination is made.

In the character processing method according to the ninth aspect of the present invention, a phonetic input step of inputting a phonetic character string in which each character of a sentence to be created is represented by phonetic characters, and phonetic characters are input from the phonetic input step. Each time, the reading evaluation step that determines whether the input reading character corresponds to a character with a predetermined erroneous input form, and the reading character that is determined when the reading evaluation step determines that the input character is an erroneous input character Is replaced with a preset alternative reading character.

A character processing method according to the invention of claim 10 is
When a reading input step of inputting a reading character string in which each character of a sentence to be created is represented by phonetic characters and the input reading character string corresponds to a predetermined erroneous input mode, the reading character string An extended candidate generation step of generating an extended reading character string by replacing the corresponding portion with another reading character according to a predetermined rule, and generating an extended candidate by converting the extended reading character string into a notation character string; A candidate display step for displaying the expansion candidate and the expanded reading character string generated by the candidate generation step, and a candidate selection step for the operator to select a desired expansion candidate from the expansion candidates displayed in the candidate display step, A text storage space that stores the extension candidate and the extension history information indicating that the extension candidate is generated based on the extended reading character string. Position information based on the position information of the expansion candidate searched in the correction position searching step and the correction position searching step for searching the position information in the converted sentence of the expansion candidate associated with the expansion history information. And a search result display step of displaying the character displayed at the position indicated by in an identifiable form.

[0025]

In the invention of claim 1, the reading input means inputs a reading character string in which each character of the sentence to be created is represented by a phonetic character. The extension candidate generation means replaces the corresponding part in the reading character string with another reading character according to a predetermined rule when the input reading character string corresponds to a predetermined erroneous input mode. Then, while generating various extended reading character strings, the extended reading character strings are converted into notation character strings.
The display attribute updating means gives a display attribute different from that of other reading character strings to the reading character string portion replaced by the expansion candidate generating means in the extended reading character string generated by the expansion candidate generating means. For example, in the device of the embodiment, an underline display is added. And the candidate display means is
The extended reading character string is displayed according to the display attribute given by the display attribute updating means.

In the invention of claim 2, the reading input means inputs a reading character string in which each character of the sentence to be created is represented by a phonetic character. The extension candidate generation means replaces the corresponding part in the reading character string with another reading character according to a predetermined rule when the input reading character string corresponds to a predetermined erroneous input mode. Then, while generating various extended reading character strings, the input reading character string and the extended reading character string are converted into notation character strings to generate expansion candidates. The display order setting means sets the display order of the expansion candidates generated by the expansion candidate generation means according to a predetermined rule. The candidate display means displays the extended candidate and the extended reading character string according to the set display order.

In the third aspect of the present invention, the maximum likelihood candidate determination means selects the extended reading character string of the extension candidate set to the first priority output order by the display order setting means and the input reading character string. It is determined whether the extended reading character string is equal to the corresponding character string. And the warning output means is
When the maximum likelihood candidate determination means determines that the reading character string and the extended reading character string are different, a warning is output in a predetermined format to call the operator's attention.

In the invention of claim 4, the reading input means inputs a reading character string in which each character of the sentence to be created is represented by a phonetic character. Every time a reading character is input from the reading inputting unit, the reading evaluation unit determines whether or not the input reading character corresponds to a character having a predetermined erroneous input form. The reading character correction unit replaces the determined reading character with a preset alternative reading character when the reading evaluation unit determines that the input character is an erroneous input character.

In the invention of claim 5, the phonetic input means inputs a phonetic character string in which each character of the sentence to be created is represented by a phonetic character. The extension candidate generation means, when the input reading character string corresponds to a predetermined erroneous input mode, replaces the relevant portion in the reading character string with another reading character according to a predetermined rule. The expansion candidate is generated by converting the expansion reading character string into the notation character string while generating the string. The candidate display means displays the extension candidate and the extended reading character string generated by the extension candidate generating means. The candidate selecting means selects an extension candidate desired by the operator from the extension candidates displayed by the candidate displaying means. The text storage unit stores the extension candidate in association with the extension history information indicating that the extension candidate is generated based on the extended reading character string. The correction location search means searches the text storage means for position information in the conversion text of the expansion candidate associated with the expansion history information. The search result display means displays the character displayed at the position indicated by this position information in an identifiable form based on the position information of the expansion candidate searched by the correction location search means.

In the character processing method according to the sixth aspect of the present invention, first, a phonetic character string in which each character of a sentence to be created is represented by a phonetic character is input. Next, when the input reading character string corresponds to a predetermined erroneous input mode, an extended reading character string is generated by replacing the corresponding part in the reading character string with another reading character according to a predetermined rule. Along with
An extension candidate is generated by converting the extension reading character string into a notation character string. Further, in the generated extended reading character string, a display attribute different from that of other reading character strings is given to the reading character string portion replaced in the extension candidate generation step. Further, the extended reading character string is displayed according to the assigned display attribute.

In the character processing method according to the seventh aspect of the present invention, first, a phonetic character string in which each character of a sentence to be created is represented by a phonetic character is input. Next, when the input reading character string corresponds to a predetermined erroneous input mode, an extended reading character string is generated by replacing the corresponding part in the reading character string with another reading character according to a predetermined rule. Along with
An extension candidate is generated by converting the input reading character string and extended reading character string into a notation character string. further,
The display order of the generated expansion candidates is set according to a predetermined rule. Further, the extension candidate and the extended reading character string are displayed according to the set display order.

In the character processing method according to the present invention, the extended reading character string of the extension candidate set to the first priority output order in the display order setting step and the input reading character string are expanded. It is determined whether the reading character string is equal to the corresponding character string. Then, when it is determined that the read character string and the extended read character string are different, a warning is output in a predetermined format.

In the character processing method according to the ninth aspect of the present invention, first, a reading character string in which each character of a sentence to be created is represented by a phonetic character is input. Next, each time a phonetic character is input, it is determined whether the input phonetic character corresponds to a character having a predetermined erroneous input form. Then, when it is determined that the input character is an erroneous input character, the determined reading character is replaced with a preset alternative reading character.

A character processing method according to the invention of claim 10 is
First, a phonetic character string in which each character of a sentence to be created is represented by phonetic characters is input. Next, when the input reading character string corresponds to a predetermined erroneous input mode, an extended reading character string is generated by replacing the corresponding part in the reading character string with another reading character according to a predetermined rule. At the same time, an extended candidate is generated by converting the extended reading character string into a notation character string. Further, the generated expansion candidates and expanded reading character strings are displayed. Further, the operator selects a desired expansion candidate from the displayed expansion candidates.
Further, the extension candidate and the extension history information indicating that the extension candidate is generated based on the extension reading character string are stored in association with each other. Then, the position information in the converted sentence of the expansion candidate associated with the expansion history information is searched. Then, based on the position information of the searched expansion candidate, the character displayed at the position indicated by this position information is displayed in an identifiable form.

[0035]

Embodiments of the present invention will be described in detail below with reference to the drawings. First, as used in this document, "reading" means that each character of a sentence to be created, for example, a sentence mixed with kanji and kana is represented by a phonetic symbol, for example, kana. In this case, this “reading” includes punctuation marks and symbols such as parentheses that form a sentence.

Further, in this document, "input candidate" means a character candidate generated by performing Kana-Kanji conversion on an input reading character string, and "extended candidate" is obtained by correcting the input reading character string. A character candidate generated by performing Kana-Kanji conversion on the extended reading character string. Then, when collectively including both, it is simply expressed as “candidate”.

FIG. 1 shows a schematic configuration of a character processing device in an embodiment of the present invention. In FIG. 1, reference numeral 1 denotes a reading input unit, which generates a corresponding code when a "reading key" on a JIS-arranged keyboard device is pressed. In addition to the "reading key", it also has various control keys such as a "conversion key", a "confirmation key", a "number key", and a "search key".
An input reading string temporary storage unit 2 sequentially stores reading codes input from the reading input unit 1 and temporarily stores them as a series of reading strings.

Reference numeral 3 is a word dictionary which stores word reading, notation and part-of-speech information. There is an independent word dictionary and an auxiliary word dictionary of the same format, and the heading is kana reading. Reference numeral 4 denotes an extended candidate generation unit, which interprets an input reading string and an extended reading string obtained by replacing a part of the reading string with another reading as a phrase string by referring to the word dictionary 3 when the conversion key is pressed. Generate as an expansion candidate.

Reference numeral 5 denotes an expansion candidate temporary storage unit, which is used for the "interpretation reading", "notation", and expansion of the expansion candidates generated by the expansion candidate generating unit 4.
"Grammar attribute information" is temporarily stored. A display device 6 outputs and displays the designated candidates on the display device of the liquid crystal panel. Reference numeral 7 denotes an attention candidate temporary storage unit, which is an extension candidate temporary storage unit 5
The identification number of one of the expansion candidates stored in is temporarily stored.

Reference numeral 8 denotes a display attribute updating unit which, among the candidates in the extended candidate temporary storage unit 5, adds the display attribute of reverse display to the notation information corresponding to the candidate number in the attention candidate temporary storage unit 7.
Further, a display attribute of underline display is added to a different portion of the interpretation reading of the extension candidate in the extension candidate temporary storage unit 5 and the input reading of the input reading string temporary storage unit 2. Reference numeral 9 is a correction number order alignment unit, which uses the number of places in the interpretation reading of the extension candidate in the extension candidate temporary storage unit 5 different from the input reading in the input reading row temporary storage unit 2 as a sort key, and the extension candidates and the like in ascending order. Align.

Reference numeral 10 denotes a correction position order aligning unit, which is the closest to the beginning among the different readings of the extension candidates in the extension candidate temporary storage unit 5 from the input readings in the input reading sequence temporary storage unit 2. The expansion candidates are arranged in ascending order using the position from the beginning of the thing as a sort key. A likelihood calculation unit 11 calculates the total likelihood of the interpretation reading of the designated extension candidate as the score sum of the reading adjacent matrix elements prepared in advance.

A likelihood order sorting unit 12 sorts the expansion candidates and the like in descending order using the likelihood of the interpretation reading of the expansion candidates in the expansion candidate temporary storage unit 5 calculated by the likelihood calculation unit 11 as a sort key. Reference numeral 13 denotes a display order setting unit, in the expansion candidates in the expansion candidate temporary storage unit 5, the first sort key is “likelihood”, the second sort key is “correction number”, and the third sort key is “correction position”. Arrange in priority order. Also, if the likelihood of a particular candidate is significantly higher than a predetermined level, then the other candidates are omitted.

Reference numeral 14 denotes a candidate display section for displaying the expansion candidates in the expansion candidate temporary storage section 5 on the display device 6 as a menu. At this time, the display attributes are displayed according to those set by the display attribute updating unit 8. Reference numeral 15 is a maximum likelihood candidate determination unit, and when the interpretation reading of the first candidate in the extension candidate temporary storage unit 5 is different from the input reading in the input reading string temporary storage unit 2, the first candidate is an extension candidate whose error has been corrected. Then, a warning signal is output.

Reference numeral 16 is a warning output unit, which is a maximum likelihood candidate determination unit 15
When the warning signal is received from the warning message, the warning message is output to the candidate display unit 14, and the warning message is displayed at the lower end of the screen of the display device 6. At the same time, an alarm that sounds a warning sounds. Reference numeral 17 denotes a text storage unit which stores the notations of the input candidates and the selected expansion candidates and their correction history information in association with each other. The correction history information holds information indicating that the interpretation reading is not the input reading.

A candidate selection unit 18 sequentially increments the attention sequence number on the attention candidate temporary storage unit 7 when the "conversion key" on the keyboard device is pressed. Then, if the current number is the same as the number of candidates (the maximum candidate ordinal number), the next is set to "1". When a specific candidate is designated by the “number key”, the corresponding candidate is stored in the text storage unit 17. When the “confirmation key” and the “non-conversion key” are pressed, the expansion candidate corresponding to the attention candidate number in the attention candidate temporary storage unit 7 at that time is stored in the text storage unit 17.
Further, when the "reading key" is input, the extended candidate corresponding to the attention candidate number in the attention candidate temporary storage unit 7 at that time is stored in the text storage unit 17, and the input reading is input and read. Write to the column temporary storage unit 2.

Reference numeral 19 denotes a correction location search unit that searches the correction history information in the text storage unit 17 according to the keystroke of the "search key", and the position information of the corresponding text in the text storage unit 17, for example, conversion. Information such as the position of the character from the beginning of the character string of the word is stored in the search result temporary storage unit 20. A search result temporary storage unit 20 temporarily stores the position information of the candidate with the correction history searched by the correction location searching unit 19.

Reference numeral 21 denotes a search result display section, which displays the display attribute of the character at the corresponding position on the display device 6 in "reverse" based on the position information stored in the search result temporary storage section 20.
Reference numeral 22 denotes a reading evaluation unit, which reads the reading sequence in the input reading sequence temporary storage unit 2 every time the reading in the input reading sequence temporary storage unit 2 changes in accordance with the reading input from the reading input unit 1. The validity of reading is evaluated based on the syllable structure and the pronunciation characteristics, and when the input character is impossible or the appearance frequency is extremely low, the reject signal and the alternative candidate reading string are output.

Reference numeral 23 is a reading correction unit, which is a reading evaluation unit 22.
When a reject signal is output from, the reading string in the input reading string temporary storage unit 2 is replaced with the alternative candidate reading string. Incidentally. In addition to this, the actual document creating apparatus requires a keyboard device for reading, entering numbers, and entering various control keys, and various constituent elements for correcting segment breaks and text editing. The detailed description of is omitted.

Further, the relationship between the constituent elements in the figure is explained mainly along the flow of data, and the exchange of control signals between the constituent elements not shown in the figure will be described in the following description. Next, the operation of the character processing device in this embodiment will be described. FIG. 2 is a flowchart showing the basic operation of the character processing device according to this embodiment. 3 to 6 are diagrams schematically showing display examples of the display device in each step of the basic operation shown in FIG. 2, which will be appropriately referred to in the following description. "Conversion process" First, a description will be given assuming a state in which input work has progressed to some extent and is in an input waiting state. A display example of this input waiting state is shown in FIG. In the figure, the underlined cursor position indicates the input position.

(101) Reading input processing step In the input waiting state, when the operator inputs a "reading key" from the reading input section 1, the corresponding reading codes are sequentially input to the input reading row temporary storage section 2 as an input reading row. Accumulated. Further, on the screen of the display device 6, the reading input by the keystroke before conversion is highlighted (see FIG. 4).

At this time, each time the reading in the input reading string primary storage unit 2 changes, the reading evaluation unit 22 determines the validity as a Japanese reading for the stored reading string. For example, as shown in FIG. 4, if the user erroneously inputs the semi-voiced point “°” after reading “U”, the reading evaluation unit 22 is appropriate for the semi-voiced point “°”. Decide if it is not. In the case of this example, the semi-dakuten “°” is judged to be an impossible character due to the Japanese syllable characteristics. And
The reading correction unit 23 automatically deletes this semi-voiced point "°". Subsequently, the operator key-inputs the character string shown in FIG.

In this state, the operator confirms the reading input on the display on the display device 6. Then, when the "conversion key" is pressed with the intention that the reading of the input reading string temporary storage unit 2 is written in kanji and kana, the control is transferred to (102). Otherwise, it returns to the input waiting state. (102) Extended conversion candidate generation step The extended candidate generation unit 4 interprets a part of the reading of the input reading string temporary storage unit 2 into another reading, thereby performing Kana-Kanji conversion with some input errors allowed. Here, as the reading conversion method, the same method as in the conventional case is used. For example, "ぢ"
An error-prone set of pseudonyms, such as "ji", "zu" and "zu", and "to" and "to" is stored in the replacement table. Then, if the input reading matches the one stored in this reading table, the reading in the input reading row is replaced with the corresponding reading in the reading table. Then, kana-kanji conversion is performed on the replaced reading character string (extended reading character string). Further, the conversion result is stored in the extension candidate temporary storage unit 5.

When the "conversion key" is pressed, the reverse-displayed reading of the input position is converted into a kanji-kana mixed string for the replacement character string generated in consideration of an input error, and the expansion candidate is displayed. (103) Display Order Setting Step The display order setting unit 13 performs a display order setting process on the candidates in the extended candidate temporary storage unit 5, and sets the first sort key to “likelihood” and the second sort key to “correction number. , And the third sort key is set as “correction position” in order of priority. Details of the display order setting process will be described below with reference to FIG. "Display Order Setting Process" First, the display order setting unit 13 delivers the extension candidate character string stored in the extension candidate primary storage unit 5 to the likelihood calculation unit 11.

(201) Likelihood calculation step The likelihood calculation unit 11 obtains, for each extended read of the extended candidate in the extended candidate temporary storage unit 5, the sum of scores of the reading adjacency frequency matrix obtained from the connection of the readings. For example, in the case of reading “koso rou”, the total likelihood S is obtained using an adjacent frequency matrix having a score of 5 levels as shown in FIG. For example, "ko → u""u → shi""shi → yo""yo →
The sum of the evaluation values of the four reading transitions of "u" is 5 + 3 + 3 +
4 = 15 points are counted, and in order to compare with a character string having a different length, the average value of reading transitions is further obtained, and the fractional part is truncated. "Let's do this" with the value calculated in this way
The likelihood of In the case of this example, since 15/4 = 3.75, "3" is the likelihood.

(202) Likelihood Order Alignment Step When the likelihood calculation unit 11 assigns a likelihood to each different reading of candidates in the extended candidate temporary storage unit 5, the likelihood order alignment unit 12
The candidates are rearranged in descending order of likelihood. FIG. 11 shows the order of candidates in the expansion candidate temporary storage unit 5 when rearranged. (203) Correction Point Number Alignment Step Next, the correction number order arranging unit 9 determines that the candidates in the expanded candidate temporary storage unit 5 that have been subjected to the likelihood order arrangement have the same number of correction points among the candidates of the same likelihood. Sort in ascending order. The result of this rearrangement is shown in FIG.

(204) Correction Position Order Alignment Step Further, the likelihood order alignment unit 12 sets the candidates in the expansion candidate temporary storage unit 5 which have been ordered in the correction position number as candidates of the same likelihood and the same correction place. On the other hand, the first correction position is rearranged in the order of being closer to the beginning. The result of this rearrangement is shown in FIG. In addition, the candidate number of the highest priority candidate at the top as a result of being arranged in the priority order is recorded in the focused candidate temporary storage unit 7.

When the above process is completed, the process returns to the process shown in FIG. (104) Display Attribute Setting Step The display attribute updating unit 8 attaches a reverse attribute to the candidate set at the beginning in the extended candidate temporary storage unit 5. Further, in the interpretation reading of each extension candidate, the portion interpreted as another reading is given an underlined attribute.

(105) Conversion Result Display Step The candidate display unit 14 displays the conversion result in the expansion candidate temporary storage unit 5 on the display device 6 based on the display attribute information. The display state is shown in FIG. In the figure, the first candidate and the second candidate are displayed in order from the top, and the first candidate is highlighted in reverse meaning the current candidate of interest. In addition, corresponding interpretation readings are simultaneously displayed for each extension candidate, and portions different from the input readings are underlined.

(106) Maximum-Likelihood Candidate Determination Step The maximum-likelihood candidate determination unit 15 compares the reading sequence of the first candidate in the extended candidate temporary storage unit 5 with the input reading sequence in the input reading sequence temporary storage unit 2. If they do not match, the control is moved to (107), and if they match, the control is moved to (108). (107) Warning Output Step When it is determined that the reading sequence of the first candidate and the input reading sequence do not match, the maximum likelihood candidate determination unit 15 outputs a warning signal to the warning output unit 16. The warning output unit 16 receives the warning signal, outputs the warning sentence to the candidate display unit 14, and causes the display device 6 to display the warning sentence. Further, the warning output unit 16 sounds an alarm expressing a warning.

In the example shown in FIG. 6, the interpretation reading "Koushou" of the first candidate is different from the input reading "Kousou", so "Kana-Kanji conversion" was performed at the bottom of the screen.
Please be careful ”warning message is displayed and the alarm emits a beep. (108) Candidate Selection Step The operator selects a desired expansion candidate from the expansion candidates displayed on the display device 6 and inputs the corresponding candidate number from the keyboard device. Alternatively, the "conversion key" is typed, the current candidate of interest is sequentially updated, the desired expansion candidate is selected, and the "decision key" and "reading key" are typed,
Select attention candidates. Details of this candidate selection process are shown in FIG.
Will be described below. "Candidate Selection Process" (301) Input Judgment Step When the operator presses a control key from the keyboard while the expansion candidate is displayed on the display device 6, the process branches as follows depending on the difference of the control key. For example, when the "reading key" is input by keystroke, the process proceeds to (302). When the "confirmation key" is entered, the process proceeds to (303). When the "conversion key" is entered, the process proceeds to (304). When the "number key" is entered, the process proceeds to (305).

(302) Reading Key Processing Step When a new reading is entered from the reading input section 1, a correction history code indicating the presence or absence of correction is attached to the attention candidate being displayed and stored in the text storage section 17, After clearing the contents of the input read string temporary storage unit 2, the extension candidate temporary storage unit 5, and the other conversion-in-progress information storage units, after storing the corresponding readings in the input read string temporary storage unit 2, it returns to the input waiting state. .

(303) Confirmation key processing step When the "confirmation key" is pressed, a correction history code indicating the presence or absence of correction is attached to the attention candidate being displayed and the text storage unit 1 is added.
After storing in 7, the contents of the input reading string temporary storage unit 2, the extension candidate temporary storage unit 5, and other conversion in-progress information storage units are cleared, and the process returns to the input waiting state. Like the "Enter key"
Also in the case of the "non-conversion key", it is considered that the reading hiragana notation of the attention candidate is selected, and the same processing as in the case of the "confirmation key" is performed.

(304) Conversion key processing step The "conversion key" in the candidate display state is an instruction to display a homophone in the next priority order, and the candidate selection unit 18 causes the candidate selection section 18 to temporarily display the candidate candidate in response to the keystroke of the "conversion key". The candidate number in the storage unit 7 is incremented. Then, the candidate display unit 14 displays the corresponding candidate on the display device 6 accordingly. After that, the process returns to the input waiting state again.

(305) Candidate No. Processing Step When a number corresponding to an expansion candidate on the display device 6 is input, the candidate selection unit 18 adds a correction history code indicating correction or not to the corresponding expansion candidate and stores the text. After storing in the unit 17, the contents of the input read string temporary storage unit 2, the extension candidate temporary storage unit 5, and other conversion in-progress information storage units are cleared, and the state returns to the input waiting state.

In this state, the operator can display the corrected portion in order to confirm whether the character input is correct. Hereinafter, the correction location search process will be described with reference to FIG. "Correction point search process" (401) Correction point search instruction step After the input is completed once, the operator can automatically check whether or not the input reading is corrected by the "search key".
Type the key.

(402) Correction Point Retrieval Step The correction point searching unit 19 searches the text storage unit 17 for correction history information recorded in association with the text (notation of a character string) in order from the beginning of the text. Then, when a text having information indicating that the input reading is corrected is searched, the position information of the searched text is temporarily stored in the search result temporary storage unit 20.

(403) Correction Point Display Step The search result display section 21 displays the position of the corresponding text on the display screen on the display device 6 with the inversion attribute based on the position information in the search result temporary storage section 20. To do. 7 and 8
Shows the states before and after the highlighted display of the corrected portion. (Modification) In the reading input process (101) of the above embodiment,
The case of automatically deleting an erroneously input character has been described. Here, an operation of correcting an erroneously input character will be described with reference to FIGS. 16 to 18. "Sequential correction process" (501) Reading character input step When the reading character is further input from the reading input unit 1 while the input reading string is displayed on the display device 6, the reading is temporarily stored in the input reading string. Sequentially stored in the unit 2.

(502) Reading Evaluation Step The reading evaluation unit 22 targets the reading string in the input reading string temporary storage unit 2 every time a reading character is input from the reading input unit 1 and determines the validity of reading of Japanese. Is evaluated based on syllable structure and pronunciation characteristics. Then, when it is impossible or when the appearance frequency is extremely low, the reject signal and the alternative candidate reading are output.

(503) Reading correction step When the reading correction section 23 receives the reject signal and the alternative candidate reading from the reading evaluation section 22, the reading correction section 23 replaces the reading row of the input reading row temporary storage section 2 with the display device at the same time. A warning display indicating that the automatic reading correction has been performed is displayed in step 6. This example is shown in FIG.
And shown in FIG.

As described above, by automatically correcting the reading character string and displaying the warning for the character string that is likely to be erroneously input, the operator can save the trouble of erroneous input processing and It can be done efficiently. Further, even when the automatic correction is inappropriate, it is possible to prevent the correction portion from being overlooked by the warning display, and the inappropriate correction portion is overlooked.

As described above, in the character processing apparatus according to the above embodiment, when an erroneous input is made at the time of inputting Japanese, correction candidates are automatically generated and the difference from the original reading is explicitly displayed. Therefore, the visibility of the correction candidate can be increased. Further, by issuing a warning when the first candidate is set as a correction candidate, the candidate confirmation work becomes easy.
In addition, after the input is completed, the corrected portion can be searched and confirmed again. Furthermore, when a character or the like that is remarkably inappropriate as Japanese is input, it can be determined and automatically corrected without waiting for the conversion instruction. With such a function, the operability of character input work is significantly improved.

In the above embodiment, both the character notation and the interpretation reading are displayed as the expansion candidates, but the character notation may not be displayed. Furthermore, although the warning output unit 16 is configured to display a warning sentence in the above-described embodiment, it may be configured to perform another warning display such as screen flashing or blinking of some characters. .

[0073]

As described above, in the invention according to claim 1 and claim 6, the portion which is interpreted as a reading different from the input reading sequence is displayed by using a display form different from the others. Therefore, the visibility of the corrected portion is improved.
Further, in the invention according to claim 2 and claim 7,
It is configured to interpret the input reading character string into different readings, generate an extended reading character string, and further prioritize and display the extended candidates generated based on this extended reading character string. Therefore, it is possible to easily select a desired character candidate when inputting.

Further, in the inventions of claims 3 and 8, a warning is issued when an extended candidate generated based on an extended reading character string different from the input reading character string is displayed as the maximum likelihood candidate. Is configured to emit,
It is possible to prompt the operator to confirm whether or not the expansion candidate is generated. Further, in the inventions of claims 4 and 9, when a character or the like that is less likely to be used grammatically or commonly is input, the erroneous input character is corrected according to a preset rule. Since it is configured, an appropriate reading input can be performed without the operator recognizing, and the operator's operation load is reduced.

Further, in the inventions of claims 5 and 10, after a desired character is selected from the expansion candidates generated by correcting the input reading character string to generate a sentence, the character in the sentence is selected. Among these, the character part generated based on the corrected reading character string is configured to be displayed separately from other parts. Can be easily confirmed or corrected.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a character processing device according to an embodiment of the present invention.

FIG. 2 is a flowchart showing a basic operation of the character processing device shown in FIG.

FIG. 3 is an explanatory diagram showing an example of a screen display of a reading character string input to the character processing device shown in FIG. 1.

FIG. 4 is an explanatory diagram showing a screen display example during an input operation of the character processing device shown in FIG. 1.

5 is an explanatory diagram showing an example of a screen display during input work of the character processing device shown in FIG. 1. FIG.

6 is an explanatory diagram showing a screen display example during an inputting operation of the character processing device shown in FIG. 1. FIG.

FIG. 7 is an explanatory diagram showing another example of the surface display according to the embodiment of the present invention.

FIG. 8 is an explanatory diagram showing a state of screen display during reading evaluation and reading correction operations in the embodiment of the present invention.

FIG. 9 is a flowchart showing a display order setting process of the character processing device according to the embodiment of the present invention.

FIG. 10 is a schematic diagram showing an adjacency matrix of a likelihood calculator of the character processing device according to the embodiment of the present invention.

FIG. 11 is an explanatory diagram showing how the display order is updated in the display order setting unit of the character processing device according to the embodiment of the present invention.

FIG. 12 is an explanatory diagram showing a state of updating the display order in the display order setting unit of the character processing device according to the embodiment of the present invention.

FIG. 13 is an explanatory diagram showing a state of updating the display order in the display order setting unit of the character processing device according to the embodiment of the present invention.

FIG. 14 is a flowchart showing an operation of a candidate selection process of the character processing device according to the embodiment of the present invention.

FIG. 15 is a flowchart showing an operation of a correction portion search process of the character processing device according to the embodiment of the present invention.

FIG. 16 is a flowchart showing the operation of the successive correction processing of the character processing device according to the embodiment of the present invention.

FIG. 17 is a schematic diagram showing an example of a screen display in the successive correction processing of the character processing device according to the embodiment of the present invention.

FIG. 18 is a schematic diagram showing an example of a screen display after correction in the successive correction processing of the character processing device according to the embodiment of the present invention.

FIG. 19 is a block diagram showing a configuration of a conventional character processing device.

FIG. 20 is a schematic diagram showing an example of a screen display of a conventional character processing device.

FIG. 21 is a schematic diagram showing another example of screen display of a conventional character processing device.

[Explanation of symbols]

 1 Reading Input Unit 4 Extended Candidate Generation Unit 6 Display Device 8 Display Attribute Update Unit 9 Correction Number Order Alignment Unit 10 Corrected Position Order Alignment Unit 11 Likelihood Calculation Unit 12 Likelihood Order Alignment Unit 13 Display Order Setting Unit 14 Candidate Display Unit 15 Maximum likelihood candidate determination unit 16 Warning output unit 18 Candidate selection unit 19 Corrected portion search unit 21 Search result display unit 22 Reading evaluation unit 23 Reading correction unit

─────────────────────────────────────────────────── ───

[Procedure amendment]

[Submission date] February 17, 1994

[Procedure Amendment 1]

[Document name to be amended] Statement

[Name of item to be amended] Claims

[Correction method] Change

[Correction content]

[Claims]

[Procedure Amendment 2]

[Document name to be amended] Statement

[Name of item to be amended] Detailed explanation of the invention

[Correction method] Change

[Correction content]

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character processing device and a character processing method which are used in a word processing system such as a word processor or a personal computer and which is controlled by a digital computer and have a function of correcting an erroneous input of a Japanese input document.

[0002]

2. Description of the Related Art In recent years, a kana-kanji conversion method has become the mainstream as a Japanese input method. This method uses the reading of each character of the document as input data to the document processing apparatus to specify the notation of the sentence to be created. Then, when creating a document in which kana-kanji conversion processing uses a phonetic character string as input data, there are many erroneous instructions due to typos, incorrect kana usage, shift omission, etc., when inputting and instructing this reading. .

Therefore, conventionally, a kana-kanji conversion method has been devised for obtaining a desired kanji-kana mixed string even with such an incorrect reading input. FIG. 19 is a block diagram showing the configuration of a conventional character processing device having such an erroneous input correction function. In FIG. 2, reference numeral 1 is a reading input unit, which generates a code corresponding to a character keyed on the keyboard device.

Reference numeral 2 denotes an input reading string temporary storage unit which sequentially stores the character codes input from the reading input unit 1 and temporarily stores them as a series of character reading strings. Reference numeral 3 is a replacement table for storing a set of pseudonyms used in the semiconductor storage device, which are prone to errors. For example, "ji" and "ji", "zu" and "zu",
A group such as "to" and "to" is stored.

Reference numeral 4 is a replacement string generation unit, which searches the replacement table 3 for the readings stored in the input reading string temporary storage unit 2. Then, if there is a match, the retrieved reading in the input reading string is replaced with the corresponding reading.
A read sequence temporary storage unit 5 includes a read sequence generation unit 4
The temporary replacement result is temporarily stored as a replacement string.

A display device 6 is provided on the liquid crystal display device.
Display the character corresponding to the specified code. Reference numeral 7 denotes an attention candidate temporary storage unit that temporarily stores the sequence number of the conversion candidate currently being noticed among the plurality of conversion candidates. Reference numeral 8 denotes a conversion candidate generation unit, and when the "conversion key" on the keyboard is pressed, the input segment and the transposition column in the input reading sequence temporary storage unit 2 and the substitution sequence temporary storage unit 5 are separated into each phrase and Estimate conversion candidates in bunsetsu and generate candidates for kanji and kana mixed sequences with high probability.

A conversion candidate temporary storage unit 9 temporarily stores the conversion candidates generated by the conversion candidate generation unit 8. For example,
When a replacement sequence of "Koori" is generated for the input reading of "Kouri", "Retail (Kouri),
"Axiom (Kouri), high profit (Kouri), utilitarianism (Kouri), ice (Koori), mulberry (Koori), freeze (Koori)" is stored in association with each conversion candidate.

A candidate display unit 10 outputs the conversion candidates in the conversion candidate temporary storage unit 9 to the display device 6 and displays them on the liquid crystal screen. At this time, the attention candidate temporary storage unit 7 is referred to and the currently noticed transformation candidate among the plurality of transformation candidates is highlighted. If the contents of the attention candidate temporary storage unit 7 change, the notation of the conversion candidate corresponding to the conversion candidate number is displayed in reverse video.

Reference numeral 11 denotes a candidate selection unit, which is a candidate candidate temporary storage unit 7 when the "conversion key" on the keyboard device is pressed.
The above noted sequence number is sequentially incremented. If the current number is the same as the number of homophone candidates (maximum sequence number), then set to "1". Further, when there is a candidate instruction by keystroke of the "number key" or a target candidate instruction by keystroke of the "confirmation key" and "reading key", the corresponding candidate is determined to be the desired one.

Next, the operation of the above-mentioned conventional character processing device will be described. First, the reading input from the reading input unit 1 is temporarily stored in the input reading string temporary storage unit 2.
When the user presses the "conversion key", the reading sequence generating unit 4 refers to the reading table 3 to generate a reading sequence in which the partial reading sequence in the input reading sequence temporary storage unit 2 is replaced, and the reading sequence is generated. It is stored in the temporary storage unit 5.

After that, the conversion candidate generation unit 8 generates conversion candidates for the input reading sequence in the input reading sequence temporary storage unit 2 and the reading sequence in the reading sequence temporary storage unit 5, and stores them in the conversion candidate temporary storage unit 9. Store. Further, the leading candidate number “1” of the conversion candidates is written in the attention candidate temporary storage unit.
The candidate display unit 10 displays a list of only the candidates of the candidate number specified in the attention candidate temporary storage unit 7 on the display device 6 with an inversion attribute.

When the user presses the "conversion key", the candidate selection section 11 sequentially increments the numbers in the attention candidate temporary storage section 7, and accordingly the conversion candidate displayed on the display device 6 is highlighted. Change. As described above, for example, with respect to the input reading string “Kouri”,
If "kou" is read as "koo,"
The display shown in 0 is made. Therefore, even if you input "ice" (koori) by mistake and enter "kouri" in kana, the necessary candidates for conversion will be displayed, so you can easily select it by selecting the conversion candidate. You can enter the characters.

[0013]

However, the conventional character input device has the following inconveniences from the viewpoint of simplification of erroneous input correction during the input operation by the operator. (1) In the character processing device shown in FIG. 19, if the partial reading in the input reading string matches the reading target reading string in the reading table 3, a plurality of candidates may be detected. For example, in the case that the Japanese sound and the dakuon are read as the direct sound and the pure sound respectively, specifically, when you input "Kou try", "ko" becomes "go", "shi" becomes "ji", " Because "yo" is converted to "yo", as shown in Fig. 21, it is for teachers, for negotiations, for construction, for improvement, for factories, and for factories. ), Joint venture (goshou), auspicious trade (gosho), and obstinacy (gojo) are presented as candidates. In this case, even if not only the kana-kanji character string that is the conversion result but also the reading string that has been interpreted and interpreted is displayed together, the display of the reading string is similar and it is difficult to distinguish them. Therefore, there is a high possibility that the operator will make a selection error when selecting the desired word notation. (2) In the conventional character processing device, when the user who is accustomed to the kana-kanji conversion operation selects the candidate even if the first candidate is the correction candidate, without being aware of the correction. There were many Therefore, a similar candidate that is not originally intended may be selected, and the correction candidate is displayed, which may cause a side effect of the correction that an inappropriate candidate is selected. (3) Generally, there are individual differences in the way of input operations, and those who use the sequential decision method that decides everything at the time of conversion, and those who select the word notation and check the term after completing the input for the time being Some people use confirmation methods. However, in the conventional character processing device, once the candidate is once selected and confirmed as the text, it is not possible to confirm which part the conversion candidate for the erroneous input correction is selected after the confirmation process. For this reason, when a person accustomed to the method of the post-confirmation operation "confirms the correction result", he / she must perform a difficult operation of estimating the reading from the character surface and confirming the correctness of the correction. . (4) Furthermore, in the conventional character processing device, the input reading string is difficult to exist as a reading of a Japanese word in the middle of input, such as "Fukazoku ..." or "Arkorizumu". Even if it was input, it was left as it was until the operator instructed to start "Conversion". For this reason, the operator originally had to correct the input reading at the same time as "when selecting a conversion candidate", which requires complicated operations such as punctuation and homophone selection. This increases the burden on the operator. It was

As described above, in the conventional character processing apparatus, "the visibility of the candidate is poor", "the confirmation of the correction fact is difficult", "the confirmation after confirmation is difficult", "the correction timing is late", etc.
There were various problems in order to improve the operability of the correction function for erroneously input characters. Therefore, the present invention has been made to solve the above problems, and provides a character processing device and a character processing method capable of improving the operability during correction processing of erroneous input that occurs when a reading character string is input. The purpose is to do.

[0015]

SUMMARY OF THE INVENTION The character processing apparatus according to the invention of claim 1 includes a read input means for inputting the readings string,
If the input read character string corresponding to the aspect of predetermined erroneous input, expanding that generates a read character string extended by replacing the separate read character according to the rules determined in advance that portion in read character string A display that gives a display attribute different from that of other reading character strings to the reading character string portion replaced by the expansion candidate generating means in the extended candidate generating means and the extended reading character string generated by the extended candidate generating means. It has attribute updating means and candidate display means for displaying the extended reading character string according to the display attribute given by the display attribute updating means.

According to a second aspect of the present invention, there is provided a character processing device comprising : a reading input means for inputting a reading character string; and a case where the input reading character string corresponds to a predetermined erroneous input mode. , The extended reading character string is generated by replacing the corresponding part in the reading character string with another reading character according to a predetermined rule.
And extended candidate generating means for generating an extended candidate by the Turkish, expanded weather display order setting means for setting in accordance with a predetermined rule display order of the expansion candidates generated by the extended candidate generating unit, according to the display order set And a candidate display means for displaying the complement .

The character processing apparatus according to the invention of claim 3 is different from the invention of claim 2 in that the extended reading character string of the extension candidate set to the first priority output order by the display order setting means is input. Among the reading character strings, the maximum likelihood candidate determining means for determining whether or not the character string corresponding to the extended reading character string is equal, and the maximum likelihood candidate determining means determines that the reading character string and the extended reading character string are different. When the determination is made, it has a warning output means for outputting a warning in a predetermined format.

According to a fourth aspect of the present invention, there is provided the character processing device , wherein the reading input means for inputting the reading character string and the reading character input is inputted each time the reading character is input from the reading input means. A reading evaluator that determines whether or not a character has a predetermined erroneous input form, and if the reading evaluator determines that the character is an erroneous input character, replaces the determined phonetic character with a preset alternative reading character. And a reading character correction means.

According to a fifth aspect of the present invention, there is provided a character processing device comprising : a reading input means for inputting a reading character string; and a case where the inputted reading character string corresponds to a predetermined erroneous input mode. An extension that generates an extended reading character string by replacing the relevant part of the reading character string with another reading character according to a predetermined rule, and at the same time generates an expansion candidate by converting the extended reading character string into a notation character string A candidate generation means, a candidate display means for displaying the extended candidate and the extended reading character string generated by the extended candidate generation means, and an operator for selecting a desired extended candidate from the extended candidates displayed by the candidate display means. A candidate selecting unit, an expansion candidate, a text storage unit that stores the expansion candidate in association with expansion history information indicating that the expansion candidate is generated based on the expansion reading character string, and a text storage unit. Based on the position information of the expansion candidate searched for the position information in the converted sentence of the expansion candidate associated with the expansion history information and the position information of the expansion candidate searched by the correction position searching unit, the position indicated by this position information And a search result display means for displaying the characters displayed in <1> in an identifiable form.

According to a sixth aspect of the present invention, there is provided a character processing method comprising : a reading input step for inputting a reading character string; and a case where the input reading character string corresponds to a predetermined erroneous input mode. ,
And extended candidate generation step that generates a read character string extended by replacing the separate read character according to a predetermined rule that portion of the string to read, among the extended read character string generated in the extended candidate generating step , A display attribute updating step of giving a display attribute different from other reading character string portions to the reading character string portion replaced in the extended candidate generation step, and the extended reading character string according to the display attribute given in the display attribute updating step And a candidate display step for displaying.

According to a seventh aspect of the present invention, there is provided a character processing method comprising : a reading input step for inputting a reading character string; and a case where the input reading character string corresponds to a predetermined erroneous input mode. ,
And extended candidate generating step of generating extended candidate by the Turkish generates a read character string extended by replacing the separate read character that portion of the string according to a predetermined rule reading, is generated in the extended candidate generating step and it has a display order setting step for setting in accordance with a predetermined rule display order of the extended candidate, and the candidate display step of displaying the extended candidate according to the display order set.

The character processing method according to the invention of claim 8 is different from the invention of claim 7, in that the extended reading character string of the extension candidate set to the first priority output order by the display order setting step is input. Among the read character strings, the maximum likelihood candidate determination step of determining whether or not the character string corresponding to the extended read character string is equal, and the read character string and the extended read character string are different in the maximum likelihood candidate determination step. And a warning output step for outputting a warning in a predetermined format when the determination is made.

According to a ninth aspect of the present invention, there is provided a character processing method , wherein a reading input step of inputting a reading character string and a reading character input every time the reading character is input from the reading input step. A reading evaluation step of determining whether or not a character having a predetermined erroneous input form is determined, and when the reading evaluation step determines an erroneous input character, the determined reading character is replaced with a preset alternative reading character. And a reading character correction step.

The character processing method according to the invention of claim 10,
And reading an input step of inputting readings string, if the input read character string corresponding to the aspect of predetermined erroneous input, a different read character according to a predetermined rule that portion in read character string An extended candidate generation step of generating an extended reading character string by replacing the extended reading character string and an extended candidate by converting the extended reading character string into a notation character string, and an extended candidate and an extended reading character generated by the extended candidate generation step. A candidate display step of displaying a column, a candidate selection step for the operator to select a desired expansion candidate from the expansion candidates displayed in the candidate display step, the expansion candidate, and the expansion candidate based on the expanded reading character string. The text storage step of storing the extended history information in association with the extended history information indicating that the extended history information is associated with the extended history information. Based on the correction point search step for searching the position information in the complementary converted sentence and the position information of the extension candidate searched in the correction point search step, the character displayed at the position indicated by this position information can be identified. And a search result display step for displaying.

[0025]

In the invention of claim 1, the reading input means inputs a reading character string. The extension candidate generation means replaces the corresponding part in the reading character string with another reading character according to a predetermined rule when the input reading character string corresponds to a predetermined erroneous input mode. Then, it generates a variety of extended reading string. The display attribute updating means gives a display attribute different from that of other reading character strings to the reading character string portion replaced by the expansion candidate generating means in the extended reading character string generated by the expansion candidate generating means. For example, in the device of the embodiment, an underline display is added. Then, the candidate display means displays the extended reading character string according to the display attribute given by the display attribute updating means.

[0026] In the second aspect of the present invention, reading input means inputs the read only string. The extension candidate generation means replaces the corresponding part in the reading character string with another reading character according to a predetermined rule when the input reading character string corresponds to a predetermined erroneous input mode. Then generates an extended candidate by the Turkish to generate a variety of extended read character string.
The display order setting means sets the display order of the expansion candidates generated by the expansion candidate generation means according to a predetermined rule. Candidate display means displays the extended candidate according to the display order set.

In the third aspect of the present invention, the maximum likelihood candidate determination means selects the extended reading character string of the extension candidate set to the first priority output order by the display order setting means and the input reading character string. It is determined whether the extended reading character string is equal to the corresponding character string. And the warning output means is
When the maximum likelihood candidate determination means determines that the reading character string and the extended reading character string are different, a warning is output in a predetermined format to call the operator's attention.

[0028] In the invention of claim 4, read input means inputs the readings string. Every time a reading character is input from the reading inputting unit, the reading evaluation unit determines whether or not the input reading character corresponds to a character having a predetermined erroneous input form. The reading character correction unit replaces the determined reading character with a preset alternative reading character when the reading evaluation unit determines that the input character is an erroneous input character.

[0029] In the fifth aspect of the present invention, reading input means inputs the read only string. The extension candidate generation means, when the input reading character string corresponds to a predetermined erroneous input mode, replaces the relevant portion in the reading character string with another reading character according to a predetermined rule. The expansion candidate is generated by converting the expansion reading character string into the notation character string while generating the string. The candidate display means displays the extension candidate and the extended reading character string generated by the extension candidate generating means. The candidate selecting means selects an extension candidate desired by the operator from the extension candidates displayed by the candidate displaying means. The text storage unit stores the extension candidate in association with the extension history information indicating that the extension candidate is generated based on the extended reading character string. The correction location search means searches the text storage means for position information in the conversion text of the expansion candidate associated with the expansion history information. The search result display means displays the character displayed at the position indicated by this position information in an identifiable form based on the position information of the expansion candidate searched by the correction location search means.

The character processing method according to the invention of claim 6, or <br/> not a inputs the readings string. Next, when the input reading character string corresponds to a predetermined erroneous input mode, an extended reading character string is generated by replacing the corresponding part in the reading character string with another reading character according to a predetermined rule. You
It Further, in the generated extended reading character string, a display attribute different from that of other reading character strings is given to the reading character string portion replaced in the extension candidate generation step.
Further, the extended reading character string is displayed according to the assigned display attribute.

The character processing method according to the invention of claim 7, or <br/> not a inputs the readings string. Next, when the input reading character string corresponds to a predetermined erroneous input mode, an extended reading character string is generated by replacing the corresponding part in the reading character string with another reading character according to a predetermined rule. you
By and the child to generate an extended candidate. Further, the display order of the generated expansion candidates is set according to a predetermined rule. Furthermore, Table <br/> view the extended candidate according to the display order set.

In the character processing method according to the present invention, the extended reading character string of the extension candidate set to the first priority output order in the display order setting step and the input reading character string are expanded. It is determined whether the reading character string is equal to the corresponding character string. Then, when it is determined that the read character string and the extended read character string are different, a warning is output in a predetermined format.

The character processing method according to the invention of claim 9, or <br/> not a inputs the readings string. Next, each time a phonetic character is input, it is determined whether the input phonetic character corresponds to a character having a predetermined erroneous input form. Then, when it is determined that the input character is an erroneous input character, the determined reading character is replaced with a preset alternative reading character.

A character processing method according to the invention of claim 10 is
First, you enter the readings string. Next, when the input reading character string corresponds to a predetermined erroneous input mode, an extended reading character string is generated by replacing the corresponding part in the reading character string with another reading character according to a predetermined rule. At the same time, an extended candidate is generated by converting the extended reading character string into a notation character string. Further, the generated expansion candidates and expanded reading character strings are displayed. Further, the operator selects a desired expansion candidate from the displayed expansion candidates. Further, the extension candidate and the extension history information indicating that the extension candidate is generated based on the extension reading character string are stored in association with each other. Then, the position information in the converted sentence of the expansion candidate associated with the expansion history information is searched. And
Based on the searched position information of the expansion candidate, the character displayed at the position indicated by this position information is displayed in an identifiable form.

[0035]

Embodiments of the present invention will be described in detail below with reference to the drawings. First of all, the "reading" as used in this document, the sentence to be created, for example, each character of kanji kana sentence, intended for example, those expressed in kana characters or Roman characters and the like. In this case, this “reading” includes punctuation marks and symbols such as parentheses that form a sentence.

Further, in this document, "input candidate" means a character candidate generated by performing Kana-Kanji conversion on an input reading character string, and "extended candidate" is obtained by correcting the input reading character string. A character candidate generated by performing Kana-Kanji conversion on the extended reading character string. Then, when collectively including both, it is simply expressed as “candidate”.

FIG. 1 shows a schematic configuration of a character processing device in an embodiment of the present invention. In FIG. 1, reference numeral 1 denotes a reading input unit, which generates a corresponding code when a "reading key" on a JIS-arranged keyboard device is pressed. In addition to the "reading key", it also has various control keys such as a "conversion key", a "confirmation key", a "number key", and a "search key".
An input reading string temporary storage unit 2 sequentially stores reading codes input from the reading input unit 1 and temporarily stores them as a series of reading strings.

Reference numeral 3 is a word dictionary which stores word reading, notation and part-of-speech information. There is an independent word dictionary and an auxiliary word dictionary of the same format, and the heading is kana reading. Reference numeral 4 denotes an extended candidate generation unit, which interprets an input reading string and an extended reading string obtained by replacing a part of the reading string with another reading as a phrase string by referring to the word dictionary 3 when the conversion key is pressed. Generate as an expansion candidate.

Reference numeral 5 denotes an expansion candidate temporary storage unit, which is used for the "interpretation reading", "notation", and expansion of the expansion candidates generated by the expansion candidate generating unit 4.
"Grammar attribute information" is temporarily stored. A display device 6 outputs and displays the designated candidates on the display device of the liquid crystal panel. Reference numeral 7 denotes an attention candidate temporary storage unit, which is an extension candidate temporary storage unit 5
The identification number of one of the expansion candidates stored in is temporarily stored.

Reference numeral 8 denotes a display attribute updating unit which, among the candidates in the extended candidate temporary storage unit 5, adds the display attribute of reverse display to the notation information corresponding to the candidate number in the attention candidate temporary storage unit 7.
Further, a display attribute of underline display is added to a different portion of the interpretation reading of the extension candidate in the extension candidate temporary storage unit 5 and the input reading of the input reading string temporary storage unit 2. Reference numeral 9 is a correction number order alignment unit, which uses the number of places in the interpretation reading of the extension candidate in the extension candidate temporary storage unit 5 different from the input reading in the input reading row temporary storage unit 2 as a sort key, and the extension candidates and the like in ascending order. Align.

Reference numeral 10 denotes a correction position order aligning unit, which is the closest to the beginning among the different readings of the extension candidates in the extension candidate temporary storage unit 5 from the input readings in the input reading sequence temporary storage unit 2. The expansion candidates are arranged in ascending order using the position from the beginning of the thing as a sort key. A likelihood calculation unit 11 calculates the total likelihood of the interpretation reading of the designated extension candidate as the score sum of the reading adjacent matrix elements prepared in advance.

A likelihood order sorting unit 12 sorts the expansion candidates and the like in descending order using the likelihood of the interpretation reading of the expansion candidates in the expansion candidate temporary storage unit 5 calculated by the likelihood calculation unit 11 as a sort key. Reference numeral 13 denotes a display order setting unit, in the expansion candidates in the expansion candidate temporary storage unit 5, the first sort key is “likelihood”, the second sort key is “correction number”, and the third sort key is “correction position”. Arrange in priority order. Also, if the likelihood of a particular candidate is significantly higher than a predetermined level, then the other candidates are omitted.

Reference numeral 14 denotes a candidate display section for displaying the expansion candidates in the expansion candidate temporary storage section 5 on the display device 6 as a menu. At this time, the display attributes are displayed according to those set by the display attribute updating unit 8. Reference numeral 15 is a maximum likelihood candidate determination unit, and when the interpretation reading of the first candidate in the extension candidate temporary storage unit 5 is different from the input reading in the input reading string temporary storage unit 2, the first candidate is an extension candidate whose error has been corrected. Then, a warning signal is output.

Reference numeral 16 is a warning output unit, which is a maximum likelihood candidate determination unit 15
When the warning signal is received from the warning message, the warning message is output to the candidate display unit 14, and the warning message is displayed at the lower end of the screen of the display device 6. At the same time, an alarm that sounds a warning sounds. Reference numeral 17 denotes a text storage unit which stores the notations of the input candidates and the selected expansion candidates and their correction history information in association with each other. The correction history information holds information indicating that the interpretation reading is not the input reading.

A candidate selection unit 18 sequentially increments the attention sequence number on the attention candidate temporary storage unit 7 when the "conversion key" on the keyboard device is pressed. Then, if the current number is the same as the number of candidates (the maximum candidate ordinal number), the next is set to "1". When a specific candidate is designated by the “number key”, the corresponding candidate is stored in the text storage unit 17. When the “confirmation key” and the “non-conversion key” are pressed, the expansion candidate corresponding to the attention candidate number in the attention candidate temporary storage unit 7 at that time is stored in the text storage unit 17.
Further, when the "reading key" is input, the extended candidate corresponding to the attention candidate number in the attention candidate temporary storage unit 7 at that time is stored in the text storage unit 17, and the input reading is input and read. Write to the column temporary storage unit 2.

Reference numeral 19 denotes a correction location search unit that searches the correction history information in the text storage unit 17 according to the keystroke of the "search key", and the position information of the corresponding text in the text storage unit 17, for example, conversion. Information such as the position of the character from the beginning of the character string of the word is stored in the search result temporary storage unit 20. A search result temporary storage unit 20 temporarily stores the position information of the candidate with the correction history searched by the correction location searching unit 19.

Reference numeral 21 denotes a search result display section, which displays the display attribute of the character at the corresponding position on the display device 6 in "reverse" based on the position information stored in the search result temporary storage section 20.
Reference numeral 22 denotes a reading evaluation unit, which reads the reading sequence in the input reading sequence temporary storage unit 2 every time the reading in the input reading sequence temporary storage unit 2 changes in accordance with the reading input from the reading input unit 1. The validity of reading is evaluated based on the syllable structure and the pronunciation characteristics, and when the input character is impossible or the appearance frequency is extremely low, the reject signal and the alternative candidate reading string are output.

Reference numeral 23 is a reading correction unit, which is a reading evaluation unit 22.
When a reject signal is output from, the reading string in the input reading string temporary storage unit 2 is replaced with the alternative candidate reading string. Incidentally. In addition to this, the actual document creating apparatus requires a keyboard device for reading, entering numbers, and entering various control keys, and various constituent elements for correcting segment breaks and text editing. The detailed description of is omitted.

Further, the relationship between the constituent elements in the figure is explained mainly along the flow of data, and the exchange of control signals between the constituent elements not shown in the figure will be described in the following description. Next, the operation of the character processing device in this embodiment will be described. FIG. 2 is a flowchart showing the basic operation of the character processing device according to this embodiment. 3 to 6 are diagrams schematically showing display examples of the display device in each step of the basic operation shown in FIG. 2, which will be appropriately referred to in the following description. "Conversion process" First, a description will be given assuming a state in which input work has progressed to some extent and is in an input waiting state. A display example of this input waiting state is shown in FIG. In the figure, the underlined cursor position indicates the input position.

(101) Reading input processing step In the input waiting state, when the operator inputs a "reading key" from the reading input section 1, the corresponding reading codes are sequentially input to the input reading row temporary storage section 2 as an input reading row. Accumulated. Further, on the screen of the display device 6, the reading input by the keystroke before conversion is highlighted (see FIG. 4).

At this time, each time the reading in the input reading string primary storage unit 2 changes, the reading evaluation unit 22 determines the validity as a Japanese reading for the stored reading string. For example, as shown in FIG. 4, if the user erroneously inputs the semi-voiced point “°” after reading “U”, the reading evaluation unit 22 is appropriate for the semi-voiced point “°”. Decide if it is not. In the case of this example, the semi-dakuten “°” is judged to be an impossible character due to the Japanese syllable characteristics. And
The reading correction unit 23 automatically deletes this semi-voiced point "°". Subsequently, the operator key-inputs the character string shown in FIG.

In this state, the operator confirms the reading input on the display on the display device 6. Then, when the "conversion key" is pressed with the intention that the reading of the input reading string temporary storage unit 2 is written in kanji and kana, the control is transferred to (102). Otherwise, it returns to the input waiting state. (102) Extended conversion candidate generation step The extended candidate generation unit 4 interprets a part of the reading of the input reading string temporary storage unit 2 into another reading, thereby performing Kana-Kanji conversion with some input errors allowed. Here, as the reading conversion method, the same method as in the conventional case is used. For example, "ぢ"
An error-prone set of pseudonyms, such as "ji", "zu" and "zu", and "to" and "to" is stored in the replacement table. Then, if the input reading matches the one stored in this reading table, the reading in the input reading row is replaced with the corresponding reading in the reading table. Then, kana-kanji conversion is performed on the replaced reading character string (extended reading character string). Further, the conversion result is stored in the extension candidate temporary storage unit 5.

When the "conversion key" is pressed, the reverse-displayed reading of the input position is converted into a kanji-kana mixed string for the replacement character string generated in consideration of an input error, and the expansion candidate is displayed. (103) Display Order Setting Step The display order setting unit 13 performs a display order setting process on the candidates in the extended candidate temporary storage unit 5, and sets the first sort key to “likelihood” and the second sort key to “correction number. , And the third sort key is set as “correction position” in order of priority. Details of the display order setting process will be described below with reference to FIG. "Display Order Setting Process" First, the display order setting unit 13 delivers the extension candidate character string stored in the extension candidate primary storage unit 5 to the likelihood calculation unit 11.

(201) Likelihood calculation step The likelihood calculation unit 11 obtains, for each extended read of the extended candidate in the extended candidate temporary storage unit 5, the sum of scores of the reading adjacency frequency matrix obtained from the connection of the readings. For example, in the case of reading “koso rou”, the total likelihood S is obtained using an adjacent frequency matrix having a score of 5 levels as shown in FIG. For example, "ko → u""u → shi""shi → yo""yo →
The sum of the evaluation values of the four reading transitions of "u" is 5 + 3 + 3 +
4 = 15 points are counted, and in order to compare with a character string having a different length, the average value of reading transitions is further obtained, and the fractional part is truncated. "Let's do this" with the value calculated in this way
The likelihood of In the case of this example, since 15/4 = 3.75, "3" is the likelihood.

(202) Likelihood Order Alignment Step When the likelihood calculation unit 11 assigns a likelihood to each different reading of candidates in the extended candidate temporary storage unit 5, the likelihood order alignment unit 12
The candidates are rearranged in descending order of likelihood. FIG. 11 shows the order of candidates in the expansion candidate temporary storage unit 5 when rearranged. (203) Correction Point Number Alignment Step Next, the correction number order arranging unit 9 determines that the candidates in the expanded candidate temporary storage unit 5 that have been subjected to the likelihood order arrangement have the same number of correction points among the candidates of the same likelihood. Sort in ascending order. The result of this rearrangement is shown in FIG.

(204) Correction Position Order Alignment Step Further, the likelihood order alignment unit 12 sets the candidates in the expansion candidate temporary storage unit 5 which have been ordered in the correction position number as candidates of the same likelihood and the same correction place. On the other hand, the first correction position is rearranged in the order of being closer to the beginning. The result of this rearrangement is shown in FIG. In addition, the candidate number of the highest priority candidate at the top as a result of being arranged in the priority order is recorded in the focused candidate temporary storage unit 7.

When the above process is completed, the process returns to the process shown in FIG. (104) Display Attribute Setting Step The display attribute updating unit 8 attaches a reverse attribute to the candidate set at the beginning in the extended candidate temporary storage unit 5. Further, in the interpretation reading of each extension candidate, the portion interpreted as another reading is given an underlined attribute.

(105) Conversion Result Display Step The candidate display unit 14 displays the conversion result in the expansion candidate temporary storage unit 5 on the display device 6 based on the display attribute information. The display state is shown in FIG. In the figure, the first candidate and the second candidate are displayed in order from the top, and the first candidate is highlighted in reverse meaning the current candidate of interest. In addition, corresponding interpretation readings are simultaneously displayed for each extension candidate, and portions different from the input readings are underlined.

(106) Maximum-Likelihood Candidate Determination Step The maximum-likelihood candidate determination unit 15 compares the reading sequence of the first candidate in the extended candidate temporary storage unit 5 with the input reading sequence in the input reading sequence temporary storage unit 2. If they do not match, the control is moved to (107), and if they match, the control is moved to (108). (107) Warning Output Step When it is determined that the reading sequence of the first candidate and the input reading sequence do not match, the maximum likelihood candidate determination unit 15 outputs a warning signal to the warning output unit 16. The warning output unit 16 receives the warning signal, outputs the warning sentence to the candidate display unit 14, and causes the display device 6 to display the warning sentence. Further, the warning output unit 16 sounds an alarm expressing a warning.

In the example shown in FIG. 6, the interpretation reading "Koushou" of the first candidate is different from the input reading "Kousou", so "Kana-Kanji conversion" was performed at the bottom of the screen.
Please be careful ”warning message is displayed and the alarm emits a beep. (108) Candidate Selection Step The operator selects a desired expansion candidate from the expansion candidates displayed on the display device 6 and inputs the corresponding candidate number from the keyboard device. Alternatively, the "conversion key" is typed, the current candidate of interest is sequentially updated, the desired expansion candidate is selected, and the "decision key" and "reading key" are typed,
Select attention candidates. Details of this candidate selection process are shown in FIG.
Will be described below. "Candidate Selection Process" (301) Input Judgment Step When the operator presses a control key from the keyboard while the expansion candidate is displayed on the display device 6, the process branches as follows depending on the difference of the control key. For example, when the "reading key" is input by keystroke, the process proceeds to (302). When the "confirmation key" is entered, the process proceeds to (303). When the "conversion key" is entered, the process proceeds to (304). When the "number key" is entered, the process proceeds to (305).

(302) Reading Key Processing Step When a new reading is entered from the reading input section 1, a correction history code indicating the presence or absence of correction is attached to the attention candidate being displayed and stored in the text storage section 17, After clearing the contents of the input read string temporary storage unit 2, the extension candidate temporary storage unit 5, and the other conversion-in-progress information storage units, after storing the corresponding readings in the input read string temporary storage unit 2, it returns to the input waiting state. .

(303) Confirmation key processing step When the "confirmation key" is pressed, a correction history code indicating the presence or absence of correction is attached to the attention candidate being displayed and the text storage unit 1 is added.
After storing in 7, the contents of the input reading string temporary storage unit 2, the extension candidate temporary storage unit 5, and other conversion in-progress information storage units are cleared, and the process returns to the input waiting state. Like the "Enter key"
Also in the case of the "non-conversion key", it is considered that the reading hiragana notation of the attention candidate is selected, and the same processing as in the case of the "confirmation key" is performed.

(304) Conversion key processing step The "conversion key" in the candidate display state is an instruction to display a homophone in the next priority order, and the candidate selection unit 18 causes the candidate selection section 18 to temporarily display the candidate candidate in response to the keystroke of the "conversion key". The candidate number in the storage unit 7 is incremented. Then, the candidate display unit 14 displays the corresponding candidate on the display device 6 accordingly. After that, the process returns to the input waiting state again.

(305) Candidate No. Processing Step When a number corresponding to an expansion candidate on the display device 6 is input, the candidate selection unit 18 adds a correction history code indicating correction or not to the corresponding expansion candidate and stores the text. After storing in the unit 17, the contents of the input read string temporary storage unit 2, the extension candidate temporary storage unit 5, and other conversion in-progress information storage units are cleared, and the state returns to the input waiting state.

In this state, the operator can display the corrected portion in order to confirm whether the character input is correct. Hereinafter, the correction location search process will be described with reference to FIG. "Correction point search process" (401) Correction point search instruction step After the input is completed once, the operator can automatically check whether or not the input reading is corrected by the "search key".
Type the key.

(402) Correction Point Retrieval Step The correction point searching unit 19 searches the text storage unit 17 for correction history information recorded in association with the text (notation of a character string) in order from the beginning of the text. Then, when a text having information indicating that the input reading is corrected is searched, the position information of the searched text is temporarily stored in the search result temporary storage unit 20.

(403) Correction Point Display Step The search result display section 21 displays the position of the corresponding text on the display screen on the display device 6 with the inversion attribute based on the position information in the search result temporary storage section 20. To do. 7 and 8
Shows the states before and after the highlighted display of the corrected portion. (Modification) In the reading input process (101) of the above embodiment,
The case of automatically deleting an erroneously input character has been described. Here, an operation of correcting an erroneously input character will be described with reference to FIGS. 16 to 18. "Sequential correction process" (501) Reading character input step When the reading character is further input from the reading input unit 1 while the input reading string is displayed on the display device 6, the reading is temporarily stored in the input reading string. Sequentially stored in the unit 2.

(502) Reading Evaluation Step The reading evaluation unit 22 targets the reading string in the input reading string temporary storage unit 2 every time a reading character is input from the reading input unit 1 and determines the validity of reading of Japanese. Is evaluated based on syllable structure and pronunciation characteristics. Then, when it is impossible or when the appearance frequency is extremely low, the reject signal and the alternative candidate reading are output.

(503) Reading correction step When the reading correction section 23 receives the reject signal and the alternative candidate reading from the reading evaluation section 22, the reading correction section 23 replaces the reading row of the input reading row temporary storage section 2 with the display device at the same time. A warning display indicating that the automatic reading correction has been performed is displayed in step 6. This example is shown in FIG.
And shown in FIG.

As described above, by automatically correcting the reading character string and displaying the warning for the character string that is likely to be erroneously input, the operator can save the trouble of erroneous input processing and It can be done efficiently. Further, even when the automatic correction is inappropriate, it is possible to prevent the correction portion from being overlooked by the warning display, and the inappropriate correction portion is overlooked.

As described above, in the character processing apparatus according to the above embodiment, when an erroneous input is made at the time of inputting Japanese, correction candidates are automatically generated and the difference from the original reading is explicitly displayed. Therefore, the visibility of the correction candidate can be increased. Further, by issuing a warning when the first candidate is set as a correction candidate, the candidate confirmation work becomes easy.
In addition, after the input is completed, the corrected portion can be searched and confirmed again. Furthermore, when a character or the like that is remarkably inappropriate as Japanese is input, it can be determined and automatically corrected without waiting for the conversion instruction. With such a function, the operability of character input work is significantly improved.

In the above embodiment, both the character notation and the interpretation reading are displayed as the expansion candidates, but the character notation may not be displayed. Furthermore, although the warning output unit 16 is configured to display a warning sentence in the above-described embodiment, it may be configured to perform another warning display such as screen flashing or blinking of some characters. .

[0073]

As described above, in the invention according to claim 1 and claim 6, the portion which is interpreted as a reading different from the input reading sequence is displayed by using a display form different from the others. Therefore, the visibility of the corrected portion is improved.
Further, in the invention according to claim 2 and claim 7,
It is configured to interpret the input reading character string into different readings, generate an extended reading character string, and further prioritize and display the extended candidates generated based on this extended reading character string. Therefore, it is possible to easily select a desired character candidate when inputting.

Further, in the inventions of claims 3 and 8, a warning is issued when an extended candidate generated based on an extended reading character string different from the input reading character string is displayed as the maximum likelihood candidate. Is configured to emit,
It is possible to prompt the operator to confirm whether or not the expansion candidate is generated. Further, in the inventions of claims 4 and 9, when a character or the like that is less likely to be used grammatically or commonly is input, the erroneous input character is corrected according to a preset rule. Since it is configured, an appropriate reading input can be performed without the operator recognizing, and the operator's operation load is reduced.

Further, in the inventions of claims 5 and 10, after a desired character is selected from the expansion candidates generated by correcting the input reading character string to generate a sentence, the character in the sentence is selected. Among these, the character part generated based on the corrected reading character string is configured to be displayed separately from other parts. Can be easily confirmed or corrected.

Claims (10)

[Claims] 1. A reading input means for inputting a reading character string in which each character of a sentence to be created is represented by phonetic characters, and when the input reading character string corresponds to a predetermined erroneous input mode. An extended reading character string is generated by replacing a corresponding portion in the reading character string with another reading character according to a predetermined rule, and an expansion candidate is generated by converting the extended reading character string into a notation character string. And a display attribute different from other reading character string portions in the reading character string portion replaced by the extension candidate generating means in the extended reading character string generated by the extension candidate generating means. And display candidate updating means for displaying the extended reading character string according to the display attribute provided by the display attribute updating means. A character processing device characterized by the following. 2. A reading input means for inputting a reading character string in which each character of a sentence to be created is represented by phonetic characters, and when the input reading character string corresponds to a predetermined erroneous input mode. , Generating an extended reading character string by replacing the corresponding portion in the reading character string with another reading character according to a predetermined rule, and converting the input reading character string and the extended reading character string into a notation character string Expansion candidate generation means for generating expansion candidates by doing, display order setting means for setting the display order of the expansion candidates generated by the expansion candidate generation means according to a predetermined rule, and according to the set display order A character processing device, comprising: the expansion candidate and a candidate display means for displaying the expansion reading character string. 3. The character processing device further includes: of the extended reading character string of the extension candidate set to the first priority output order by the display order setting means and the input reading character string. Maximum likelihood candidate determining means for determining whether or not the character string corresponding to the extended reading character string is equal, when the maximum likelihood candidate determining means determines that the reading character string and the extended reading character string are different, The character processing device according to claim 2, further comprising a warning output unit that outputs a warning in a predetermined format. 4. A phonetic input unit for inputting a phonetic character string in which each character of a sentence to be created is represented by phonetic characters, and the phonetic character input every time the phonetic character is input from the phonetic input unit. And a reading evaluator that determines whether or not the character corresponds to a character with a predetermined erroneous input form, and if the reading evaluator determines that the character is an erroneous input character, the reading character that is determined is an alternative reading character And a reading character correcting unit that replaces 5. A reading input means for inputting a reading character string in which each character of a sentence to be created is represented by a phonetic character, and when the input reading character string corresponds to a predetermined erroneous input mode. An extended reading character string is generated by replacing a corresponding portion in the reading character string with another reading character according to a predetermined rule, and an expansion candidate is generated by converting the extended reading character string into a notation character string. An expansion candidate generating unit, a candidate display unit displaying the expansion candidate and the expansion reading character string generated by the expansion candidate generating unit, and an operator's desire from among the expansion candidates displayed by the candidate displaying unit. Correspondence with candidate selection means for selecting an extension candidate, the extension candidate, and extension history information indicating that the extension candidate is generated based on the extension reading character string. A text storage unit that stores the text, a correction location search unit that searches the text storage unit for position information in the converted sentence of the extension candidate associated with the extension history information, and a correction location search unit that searches. And a search result display means for displaying the character displayed at the position indicated by the position information in a distinguishable form based on the position information of the expansion candidate. 6. A reading input step of inputting a reading character string in which each character of a sentence to be created is represented by phonetic characters, and when the input reading character string corresponds to a predetermined erroneous input mode. An extended reading character string is generated by replacing a corresponding portion in the reading character string with another reading character according to a predetermined rule, and an expansion candidate is generated by converting the extended reading character string into a notation character string. And a display attribute different from another reading character string portion in the reading character string portion replaced in the extension candidate generating step among the extended reading character string generated in the extension candidate generating step. A display attribute updating step of adding the display attribute updating step of displaying the extended reading character string according to the display attribute added in the display attribute updating step. And a supplementary display step. 7. A reading input step of inputting a reading character string in which each character of a sentence to be created is represented by a phonetic character, and when the input reading character string corresponds to a predetermined erroneous input mode. , Generating an extended reading character string by replacing the corresponding portion in the reading character string with another reading character according to a predetermined rule, and converting the input reading character string and the extended reading character string into a notation character string An expansion candidate generation step for generating an expansion candidate by doing, a display order setting step for setting the display order of the expansion candidates generated in the expansion candidate generation step according to a predetermined rule, and according to the set display order A character processing method comprising: the expansion candidate and a candidate display step of displaying the expansion reading character string. 8. The character processing method further comprises: selecting the extended reading character string of the extension candidate set to the first priority output order by the display order setting step and the input reading character string. A maximum likelihood candidate determination step of determining whether or not the character string corresponding to the extended reading character string is equal, and when it is determined that the reading character string and the extended reading character string are different in the maximum likelihood candidate determination step. 8. The character processing method according to claim 7, further comprising: a warning output step of outputting a warning in a predetermined format. 9. A reading input step of inputting a reading character string in which each character of a sentence to be created is represented by phonetic characters, and each time the reading character is input from the reading input step,
A reading evaluation step of determining whether or not the input read character corresponds to a character of a predetermined erroneous input form, and when the read evaluation step determines that the input character is an erroneous input character, the determined read character is A character processing method, comprising: a reading character correcting step of replacing with a preset alternative reading character. 10. A reading input step of inputting a reading character string in which each character of a sentence to be created is represented by phonetic characters, and when the input reading character string corresponds to a predetermined erroneous input mode. An extended reading character string is generated by replacing a corresponding portion in the reading character string with another reading character according to a predetermined rule, and an expansion candidate is generated by converting the extended reading character string into a notation character string. An extension candidate generating step, a candidate displaying step for displaying the extension candidate and the extended reading character string generated by the extension candidate generating step, and an operator's desire from among the extension candidates displayed in the candidate displaying step. A candidate selection step for selecting an extended candidate, the extended candidate, and the extended candidate generated based on the extended reading character string. A text storage step of storing the expanded history information in association with the expanded history information, a correction location searching step of searching for positional information in the converted sentence of the expansion candidate associated with the expansion history information, and the correction location And a search result display step of displaying, in an identifiable form, a character displayed at a position indicated by the position information based on the position information of the extension candidate searched in the search step. Method.