JP4995801B2 - Document analysis apparatus, document analysis program, and document analysis method - Google Patents

Description

The present invention relates to a document analysis apparatus , a document analysis program, and a document analysis method for recognizing a sentence in the same hierarchy by shaping a citation hierarchy even when a line break is automatically performed by a mail tool.

  In recent years, the penetration rate of personal computers is high, and with the establishment of the Internet environment, the exchange of information with overseas by e-mail has become increasingly popular. In addition, an increasing number of users use machine translation for e-mail translation.

  Until now, e-mail users have received manners education about writing e-mails at schools and companies, etc., and descriptions such as inserting line breaks with a certain number of characters were taken into consideration, With the spread of personal computers, the number of users who do not know such manners is increasing. As a result, a large number of e-mails can be seen in which characters are entered in the full width of the screen or written without line breaks.

  For this reason, some mail software has a function to automatically wrap (break) a specified number of digits. When this function is enabled, it becomes very difficult to read due to the line breaks explicitly and double line breaks by automatic wrapping. Also, when replying to an email, a quotation mark is inserted in the quoted portion of the original email, so that the line is broken even if the sentence is originally in the same hierarchy while the reply is repeated and the reply is repeated. It is also difficult to determine whether it is a response part.

Therefore, in order to prevent line breaks in unnatural places when displaying mail received from a PC, etc., the line break data in the body of received mails has been formatted to eliminate unnatural line breaks. There are some (see, for example, Patent Document 1). That is, in Patent Document 1, the first quotation mark of a quotation is stored in a mail sentence, and when a line break data is detected, a line break automatically inserted by mail software or a line break intentionally inserted by a user. If it is determined that it is an automatically inserted line break, the line break is deleted to make it a part of the quoted sentence. The method of improving the readability of the mail document by inserting a line feed according to the display width of the screen and inserting a quotation mark at the beginning of the sentence existing between the quotation character and the quotation terminator. Yes.
JP-A-11-184775

  However, in Patent Document 1, the data after detecting the quote symbol at the beginning of the line is inspected, and if the second or later quote symbol is detected, it is deleted. When line feed data is detected, the next character of the line feed data, that is, the bullet, is “” (space), “(” (open parenthesis), “?” (Question mark), “number”, “•”. (Inside of black), it is judged as an automatically inserted line break and the line break is deleted, and the previous character of the line break data is a punctuation mark, ")" (close bracket) If it is neither, it is determined that the line was inserted automatically and the line break is deleted.

  On the other hand, if it is determined that it is not an automatically inserted line break, a code meaning the end of citation is added, and when displaying on the screen, a line break is made at a position that matches the current screen width. In the case of a sentence corresponding to a quoted portion, the readability of the mail document is improved by adding a quote symbol after the line feed code. Therefore, there is a problem that even when the citation layers are different, they are recognized as data of the same layer. Further, since the mail body data A1 as shown below starts with a number, there is a problem in that the automatically inserted line feed before “2” is not deleted.

本発明の目的は、複数回のメールのやり取りが行われ引用階層が乱れた場合でも、自動改行データが挿入された場合の規則性を判定することにより、より正確な文の単位が認識できるようにメールの本文データを整形できる、文書解析プログラムおよび文書解析方法を提供することである。 [Mail body data A1]

It is an object of the present invention to recognize a more accurate sentence unit by determining regularity when automatic line feed data is inserted even when a mail is exchanged multiple times and the citation hierarchy is disturbed. To provide a document analysis program and a document analysis method capable of formatting mail body data.

The document analysis apparatus according to the present invention reads the mail body data input from the input device for each line feed , determines the data length for each line based on the number of characters in each line character string, and determines the number of quotation marks in each line character string. A document hierarchy determination unit that determines a citation hierarchy for each line based on the document hierarchy, and a citation hierarchy of a data length approximate maximum line that has a data length within a predetermined range of a maximum value among the data lengths determined by the document hierarchy determination unit If is greater than the citation hierarchy of the next line, the character string from which the quotation mark and line feed are removed from the character string of the maximum approximate data length line and the character from which the quotation mark and line feed are removed from the character string of the next line when the length of the coupling length of the string of binding strings and string to determine whether the predetermined range of the maximum value of the data length is within a predetermined range of the maximum value of the data length Above The same hierarchy determination unit that determines that the actual hierarchy of the data length approximate maximum line and the next line is a character string of the same hierarchy, and the data length approximation maximum that is determined to be the same hierarchy by the same hierarchy determination unit A document format that combines a line and the character string of the next line to form a combined character string, adds a quoting symbol representing the citation hierarchy of the line with the approximate maximum data length to the combined character string, and outputs it to the output device as a new line A part .

  According to the present invention, it is possible to recognize a more accurate sentence unit by determining regularity when automatic line feed data is inserted even when a plurality of mails are exchanged and the citation hierarchy is disturbed. You can format mail body data.

  FIG. 1 is a functional block diagram of a document analysis apparatus 11 according to the embodiment of the present invention, and FIG. 2 is a block configuration diagram showing a hardware configuration of the document analysis apparatus according to the embodiment of the present invention.

  In FIG. 2, the document analysis device 11 is realized by installing a software program such as a document analysis program in a general computer and executing the software program in the processor 13 of the arithmetic control device 12.

  The arithmetic control device 12 performs various arithmetic operations related to document analysis. The arithmetic control device 12 includes a processor 13 and a memory 14. A memory 14 stores a document analysis program 15 related to translation, and the processor 13 performs processing. When executed, the work area 16 is used. Calculation results and the like of the calculation control device 12 are displayed and output on the display device 18 that is the output device 17 and also output to the communication network via the communication control device 19.

  The input device 20 is used to input information to the arithmetic control device 12, and includes, for example, a mouse 21, a keyboard 22, a disk drive 23, and a communication control device 19. For example, the mouse 21 and the keyboard 22 are connected via the display device 18. Then, various commands are input to the arithmetic and control unit 12, and the keyboard 22, the disk drive 23, and the communication control unit 19 input mail body data.

  That is, the disk drive 23 inputs / outputs mail text data to / from a storage medium, and the communication control device 19 connects the document analysis device 11 to a communication network such as the Internet or a LAN. The communication control device 19 is a device such as a LAN card or a modem, and data transmitted / received to / from the communication network via the communication control device 19 is transmitted / received to / from the arithmetic control device 12 as an input signal or an output signal. Furthermore, a hard disk drive (HDD) 24 is provided for storing the calculation results of the calculation control device 12, the mail tool program, and the like.

  FIG. 1 is a functional block diagram of a document analysis apparatus 11 according to an embodiment of the present invention. Each functional block in the arithmetic and control unit 12 shown in FIG. 2 corresponds to each program constituting the document analysis program 15 described above. That is, when the processor 13 executes each program constituting the document analysis program 15, the arithmetic and control unit 12 functions as each functional block. The block of the storage device 25 corresponds to the storage area of the memory 14 and the hard disk drive 24 in the arithmetic control device 12.

  The input processing unit 26 performs an input interface with the outside, and receives mail text data and commands through the communication control device 19 such as the Internet and the input device 20 such as the keyboard 23.

  The output processing unit 27 performs an output interface with the outside, and outputs mail body data through the communication control device 19 such as the Internet and the output device 17 such as the display device 18.

  The control unit 28 controls the entire apparatus. The control unit 28 stores the text data of the mail sent from the input processing unit 26 in the data temporary storage unit 29 of the storage device 25, the document hierarchy determination unit 30, and the same hierarchy determination. The unit 31, the document shaping unit 32, and the sentence recognition processing unit 33 are controlled, and the contents of the data temporary storage unit 29 storing these calculation results are sent to the output processing unit 27 and output to the output device 17.

  The document hierarchy determination unit 30 determines the hierarchy of the body data of the mail obtained through the input processing unit 26. That is, the data length is determined for each line based on the number of characters in each line character string of the mail body data, and the citation hierarchy is determined for each line based on the number of quotation marks in each line character string. Then, the character string, the data length, and the citation hierarchy for each line are stored in the data temporary storage unit 29 of the storage device 25.

  The same hierarchy determination unit 31 determines whether the citation hierarchy of adjacent lines is the same hierarchy. This determination method will be described later.

  The document shaping unit 32 combines the character strings of the lines determined to be the same hierarchy by the same hierarchy determination unit 31 to form a combined character string, and adds a quotation mark representing the citation hierarchy to the combined character string as a new line. The data is stored in the temporary data storage unit 29 of the storage device 25.

  The sentence recognition processing unit 33 combines adjacent lines in the same hierarchy of the body data of the mail including the line newly given by the document shaping unit 32, determines a sentence at the position of the punctuation in the combined character string, A character string in which a quotation mark of the quotation hierarchy is added to the head of each sentence is created and stored in the data temporary storage unit 29 of the storage device 25.

  Next, operation | movement of Example 1 of the document analysis apparatus 11 concerning embodiment of this invention is demonstrated. FIG. 3 is a flowchart showing the operation of Example 1 of the document analysis apparatus 11 according to the embodiment of the present invention.

  When mail text data is input from the input device 20 to the arithmetic processing unit 12, the control unit 28 activates the input processing unit 26, and the input processing unit 26 reads data for each line feed of the mail text data (S201). . Assume that the following mail text data A2 is read.

このメールの本文データＡ２は、以下に示すメールの本文データＡ３がメールツールの設定により全角２０文字の位置で自動的に改行され、１度目の返信時に「はい。」が挿入され、さらに２度の転送が行われたものである。 [Mail body data A2]

In the mail body data A2, the mail body data A3 shown below is automatically broken at a position of 20 full-width characters according to the setting of the mail tool, and “Yes” is inserted at the first reply, and then twice. Has been transferred.

メールの本文データＡ２及びメールの本文データＡ３から分かるように、メールの本文データＡ２は、本来同一行であった文字が自動改行により異なる行になった上、引用階層も異なった状態になってしまったものである。 [Mail body data A3]

As can be seen from the mail body data A2 and the mail body data A3, the mail body data A2 has different characters due to the automatic line feed, and the citation hierarchy is also different. It is a fool.

The mail body data A2 read by the input processing unit 26 is stored in the data temporary storage unit 29 of the storage device 25. Table 1 shows an example of a table of mail body data A2 stored in the data temporary storage unit 29 of the storage device 25.

  As shown in Table 1, the array element number is a number assigned to each line of mail body data, and the data is a character string for each line of body data. The length is the data length (number of character strings) for each line of the body data, and the citation hierarchy is a hierarchy indicating the number of mail replies. In Table 1, since the data is stored while the mail body data A2 is read, the length and citation hierarchy are blank.

  When the mail body data A2 is read for each line feed and stored in the data temporary storage unit 29 of the storage device 25 as shown in Table 1, the control unit 28 activates the document hierarchy determination unit 30 to perform document hierarchy determination. (S202).

  FIG. 4 is a flowchart showing the processing contents of document hierarchy determination of the document hierarchy determination unit 30. The document hierarchy determination in the document hierarchy determination unit 30 is to determine the hierarchy of the mail body data.

  First, the document hierarchy determination unit 30 sets an initial value “1” to the line data designation variable i (S301), and the array number of the mail body data A2 stored in the data temporary storage unit 29 of the storage device 25 is set. The i-th data Si is acquired (S302). When i is “1”, “>>> mail tool setting is“ at the time of mail transmission ”is acquired as data S1.

Next, an initial value “1” is set in a variable L indicating the citation hierarchy, a variable N designating each element of the data Si, and a variable M indicating the position of the citation symbol array (S303). A variable L indicating the quoting hierarchy is a variable indicating the number of mail replies. A variable N for specifying each element of the data Si is a variable for specifying a position of an element (character string) of the data Si. A variable M indicating a position is a variable for designating a quotation mark. Table 2 shows an example of a table of quoted symbol arrays.

  When the variable M indicating the position of the reference symbol array is “1”, “>” is acquired as the reference symbol, and when the variable M indicating the position of the reference symbol array is “2”, the reference symbol “ | "Is acquired.

  Next, the quote symbol CM at the position indicated by the variable M indicating the position of the quote symbol array is acquired (304). Then, it is determined whether or not the character of the data Si starting from the variable N matches the quote symbol CM (S305). If not, it is determined whether the quote symbol CM is the last data (S306), and the quote symbol CM is the last one. If not, M = M + 1 is performed (S307), and the process returns to step S304. Thereby, in step S304, “|” is acquired as the quotation symbol CM. In this way, it is determined from the beginning of the data Si whether or not the character of the data Si matches each quote symbol stored in the quote symbol array shown in Table 2.

  Assume that i, L, N, and M are all “1”. The data S1 in the first row where i is “1” is (>>> the mail tool setting is “at the time of mail transmission.” In the determination of step S304, the variable M is 1, so the quotation mark CM Is “>”, and the character indicated by the data Si (S1) at the position indicated by the variable N (= 1) is “>”. Accordingly, in the determination in step S305, the character “>” of the data Si (S1) starting from the variable N (= 1) matches the quotation mark CM (=>), and the process proceeds to step S308.

  In step S308, the length of the matching reference symbol CM “>” is added to the variable N indicating the determination start position of the data Si. Thereby, the variable N is updated to the next determination start position. Further, 1 is added to the variable L indicating the citation hierarchy to obtain L + 1, and the variable L is updated to the next citation hierarchy. Note that the variable M indicating the position of the reference symbol array remains M = 1. This is because the variable M is updated in step S307. If the Nth position of the data Si is a space, 1 is added to the variable N to obtain N + 1, and the variable N is updated to the next determination start position.

  Then, returning to the process of step S304, steps S304, S305, and S308 are repeated until the character at the position of the variable N in the data Si does not match the quote symbol CM in step S305. If the character at the position of the variable N in the data Si does not match the quote symbol CM in step S305, the process proceeds to step S306, and steps S304, S305, and S308 are repeated until the quote symbol CM becomes the last data.

If it is determined in step S306 that the quotation mark CM is the last data, the process proceeds to step S309. In step S309, depending on the mail tool, a space may be inserted together with the quote symbol CM, and therefore the space is evaluated. It is determined whether or not the Nth character of the Si data is N ≠ 1, the Nth character is a space, and the N−1th character is not a space. If so, the space is treated as a quotation mark and the process proceeds to S308. move on. If the condition of step S309 is not satisfied, the data Si and the length of the data Si are stored in the data temporary storage unit 29 together with the variable L indicating the citation hierarchy (S310). As a result, as shown in Table 3, for the data S1 in the first line of the mail body data, the length of the data S1 is 40 and the citation hierarchy is 4. The length of the data S1 is indicated by the number of single-byte characters, and the citation hierarchy is indicated by the sum of the number of quotation marks CM and the space.

Next, it is determined whether or not the data Si is the last line (S311). If it is not the last line, 1 is added to the line data designation variable i (S312), and the process returns to step S302. Thus, the same processing is performed for the second and subsequent rows, and the processing is terminated when the last row is reached. At this time, the table of the mail body data A2 stored in the data temporary storage unit 29 of the storage device 25 stores the length of the data S1 and the citation hierarchy for Si in each row as shown in Table 4. The

  In this way, when the hierarchy of the mail body data is determined by the document hierarchy determination unit 30, the same hierarchy determination unit 31 determines whether the citation hierarchy of adjacent lines is the same hierarchy. In step S203 of FIG. 3, the same hierarchy determination unit 31 uses the data Si of the mail body data A2 from the mail body data A2 table shown in Table 4 stored in the data temporary storage unit 29 of the storage device 25. The maximum value n of the citation hierarchy and the maximum value MAX of the data length are acquired and stored in the data temporary storage unit 29. Further, “1” is set to the row data designation variable i, and “1” is set to the loop variable J (S203).

  Subsequently, the same hierarchy determination unit 31 acquires the i-th data Si and the quotation hierarchy Li from the mail body data A2 shown in Table 4 (S204). Then, it is determined whether or not the length of the acquired data Si is the data length approximate maximum row within the approximate range of the data length maximum value MAX. That is, the same hierarchy determination unit 31 determines whether or not the length of the data Si> MAX−correction value (S205). Here, the correction value is determined by using ± 2 of the data length maximum value MAX as an approximate value. This correction value may be set by the user. The reason why the correction value is set is, for example, that a space is inserted in a part of the citation hierarchy or a space is inserted between words in the case of English.

  When it is determined in step S205 that the length of the data Si is the data length approximate maximum row, the same hierarchy determination unit 31 determines whether or not the data Si + 1 in the i + 1th row which is the next row can be acquired. When the determination is made (S206) and the data Si + 1 can be acquired, the data Si + 1 and the citation hierarchy Li + 1 in the i + 1th row which is the next row are acquired (S207).

  Then, the same hierarchy determination unit 31 has a length obtained by removing the quote symbol portion of the data Si and the data Si + 1 in which the quote hierarchy Li of the data Si that is the data length approximate maximum row is larger than the quote hierarchy Li + 1 of the data Si + 1. It is determined whether the value obtained by adding is an approximate value of the data length maximum value MAX. That is, it is determined whether or not the lengths of Si and Si + 1 with Li> Li + 1 and the quotation marks removed are within the MAX approximate value (S208). When the condition of step S208 is satisfied, the same hierarchy determination unit 31 determines that the actual hierarchy of the data length approximate maximum line and the next line is a character string of the same hierarchy. This is because the data Si + 1 was originally a part of the data Si, but it can be determined that the line feed is inserted by automatic line feed.

  Next, the document shaping unit 32 removes the quotation mark of Si + 1 and combines it with Si (S209). That is, the quotation mark from the head of the data Si + 1 is removed and combined with the end of the data of the data Si. The quote symbol in this case also includes the first space following the quote symbol appended to the quote symbol.

  Then, the document shaping unit 32 determines whether or not the data Si + 1 is the last data (S210). If it is not the last data, i = i + 2 is set (S211), and the process returns to step S204. If the data Si is the last data, the document shaping unit 32 sets J = J + 1 (S212), and determines whether the loop variable J is equal to the maximum value n of the citation hierarchy. That is, it is determined whether J = n (S213). If J = n is not satisfied, i = 1 is set (S214), the process returns to step S204, and the processes of steps S204 to S214 are repeated.

  When it is determined in step S205 that the length of the data Si is not the data length approximate maximum row, if the determination in step 208 does not satisfy the condition in step S208, the process proceeds to step S210. On the other hand, if Si + 1 cannot be acquired in the determination in step S206, the process proceeds to step S212.

Here, when the processing of the first steps S204 to S214 is completed, the array numbers 1, 2, array numbers 4, 5, and array numbers 7, 8 of the mail body data A2 shown in Table 4 are combined, respectively. The mail body data is as shown in Table 5.

Further, by repeating the processing of S204 to S214 until the loop variable J becomes equal to the maximum value n of the citation hierarchy, the mail body data shown in Table 6 is obtained, and the citation hierarchy disturbed by the automatic line feed is shaped. .

  In this way, the document shaping unit 32 combines the maximum data length approximated line determined to be the same hierarchy and the character string of the next line into a combined character string, and a citation symbol representing the citation hierarchy of the data length approximate maximum line Is added to the combined character string and stored in the data temporary storage unit 29 of the storage device 25 as a new line.

  Subsequently, sentence recognition processing is performed on the body data of the mail in which the citation hierarchy shown in Table 6 is shaped (S215). FIG. 5 is a flowchart showing the processing content of the sentence recognition processing of the sentence recognition processing unit 33. The sentence recognition processing in the sentence recognition processing unit 33 is to combine adjacent lines in the same hierarchy in the body data of the mail including the line newly given by the document shaping unit 32 to form a character string of the sentence.

  In order to extract the data Si from each line of the text data of the document-shaped mail shown in Table 6, first, an initial value “1” is set to the line data designation variable i as shown in FIG. 5 (S401).

  Subsequently, the i-th data Si is acquired, the quotation mark of the acquired data Si is removed, and the i-th data Si is stored in the sentence recognition processing storage area of the temporary data storage unit 29. That is, the quotation mark is removed from the data Si and stored in the data temporary storage unit (S402). Then, the citation hierarchy Li of the removed citation symbol is stored in the data temporary storage unit 29 (S403). The citation hierarchy Li of the removed citation symbol is stored in the citation hierarchy storage area of the data temporary storage unit 29. Next, an initial value “1” is set to a variable p that serves as a loop counter for repeating the process from the i-th position until data having a different citation hierarchy appears (S404).

  Subsequently, it is determined whether or not the citation hierarchy of the citation symbol of the data Si + p can be acquired (S405). If the citation hierarchy of the citation symbol of the data Si + p cannot be acquired, the process ends. On the other hand, when the citation hierarchy of the citation symbol of the data Si + p is acquired, it is determined whether the citation hierarchy Li of the data Si and the citation hierarchy of the data Si + p are the same (S406). If the citation hierarchies are the same, the character string of the citation symbol of the data Si + p is deleted, a line feed code is added to the end of the data Si + p, and stored in the data temporary storage unit 29 (S407). These data are additionally stored in the sentence recognition processing storage area of the data temporary storage unit 29. Then, to obtain the next data, the variable p is set to p = p + 1 (S408), and the process returns to step S405.

If it is determined in step S406 that data Si having a different citation hierarchy appears, sentence determination processing is performed based on the data in the sentence recognition processing storage area (S409). in this case,
As shown in Table 7, the data in the sentence recognition processing storage area is mail text data in which data having the same citation hierarchy is stored.

  In the sentence determination processing in step S409, for example, a phrase such as “.” Is determined as the end of the sentence, and “Mail tool setting is... Subsequently, a sentence from “all” to “.” That appears next is taken as one sentence. Next, “above” is a sentence.

When the end is reached, a quotation mark is inserted at the head of each sentence (S410). That is, when a character string of a quotation symbol temporarily stored in the quotation hierarchy storage area of the data temporary storage unit 29 is added to the head of each sentence, mail body data as shown in Table 8 is obtained.

Then, it is determined whether or not the data Si is the last data (S411). If the data has not reached the last data, i = i + p is set to advance the variable i by p (S412), the process returns to step S402, and the data Si + p The processes of S402 to S411 are repeated. As a result, as shown in Table 9, “Yes.” Becomes the target of the sentence recognition process, and “Yes.” Is acquired as a sentence unit.

When the character string of the quote symbol temporarily stored in “Yes” is inserted, the mail body data as shown in Table 10 is obtained.

  According to Example 1 of the embodiment of the present invention, line breaks are automatically made according to the setting of the mail tool, and even when the data is the same in the writer of the mail and in a different hierarchy, it is automatically inserted. Since the layer disturbed by the line break is shaped and the sentence recognition process is performed in the same layer, the sentence recognition accuracy can be improved. As described above, the accuracy of the sentence recognition can be improved by shaping the layer disturbed by the line feed, so that the translation accuracy of the mail body data to be translated can also be improved. For example, in order to obtain a correct translation result in a translation apparatus, it is necessary to accurately recognize a sentence unit to be translated, but the request can be met.

  Next, the operation of Example 2 of the document analysis apparatus 11 according to the embodiment of the present invention will be described. FIG. 6 is a flowchart showing the operation of Example 2 of the document analysis apparatus 11 according to the embodiment of the present invention. The second embodiment is additionally provided with step S208a for determining whether or not the citation hierarchy Li is larger than the length of the data Si + 1 when it is determined that the condition of step S208 in the flowchart shown in FIG. 3 is not satisfied. It is.

  As a result, when the same hierarchy determination unit 31 does not satisfy the condition of step S208, that is, the citation hierarchy Li of the data length approximate maximum line is larger than the citation hierarchy Li + 1 of the next line, and the length of the combined character string is data When it is not within the approximate range of the maximum length, the citation hierarchy Li of the data length approximate maximum line is compared with the character string of the next line, and the citation hierarchy of the data length approximate maximum line is compared with the character string of the next line. When it is larger, the actual hierarchy of the data length approximate maximum line and the next line is determined to be a character string in the same hierarchy.

  Now, description will be made using the following mail body data A4 as mail body data.

[Mail text data A4]

In the mail body data A4, the data in which the line is explicitly inserted at the 11th character is automatically folded back to 20 characters according to the setting of the mail tool by exchanging the mail several times. It has become. The mail text data A4 becomes the mail text data shown in Table 11 by the processing up to step S208 in FIG.

  For the mail body data in Table 11, in step S208 of FIG. 6, the data obtained by combining “setting of mail tool” obtained by removing the quotation mark of the first data S1 and “ga” of the second data S2. The length is 20. Therefore, in step S208, it is not determined that the approximate value is 39, which is the maximum data length value MAX, and it is determined that the first data S1 and the second data S2 are not combined.

  Therefore, in the second embodiment, when it is determined in step S208 in FIG. 6 that they are not combined, the length of the second data S2 is compared with the citation hierarchy L1 of the first data S1, and the second data S2 When the length is smaller than the citation hierarchy L1 of the first data S1, it is determined that the data is in the same hierarchy (S208a). Then, the document shaping unit 32 combines the data Si and the data Si + 1 based on the determination in step S208a.

In the flowchart of FIG. 6, the mail text data shown in Table 12 is obtained by the first processing in steps S204 to S214.

Then, the mail body data shown in Table 13 is obtained by the second processing of steps S204 to S214.

  Since the data hierarchy is the same as the mail body data shown in Table 13 in the second processing of S204 to S214, the variable J reaches the maximum value n of the citation hierarchy in step S213 in FIG. If the result is the same as the previous time, the process of the document shaping unit 32 is terminated (S213), and the process proceeds to the sentence recognition process (S215).

  In the above description, when the condition of step S208 is not satisfied, the same hierarchy determination unit 31 compares the citation hierarchy Li of the data length approximate maximum line with the character string of the next line, and quotes the data length approximate maximum line. When the hierarchy is larger than the character string of the next line, it is determined that the actual hierarchy of the data length approximate maximum line and the next line are the same hierarchy character string. Compare the citation hierarchy with a predetermined citation hierarchy threshold, and if the citation hierarchy of the data length approximate maximum line is equal to or greater than the predetermined citation hierarchy threshold, the actual hierarchy of the data length approximate maximum line and the next line May be determined to be character strings in the same hierarchy. In addition, the character string of the next line of the data length approximation maximum line is compared with the predetermined number of characters, and if the character string of the next line of the data length approximation maximum line is less than the predetermined number of characters, the data length approximation You may make it determine with the actual hierarchy of the largest line and the following line being the character string of the same hierarchy.

  According to Example 2 of the embodiment of the present invention, in addition to the effect of Example 1, even when the citation hierarchy is deep, it is possible to shape a hierarchy that is disturbed by a line feed that is automatically inserted.

  Next, another example of the document analysis apparatus 11 according to the embodiment of the present invention will be described. FIG. 7 is a functional block diagram of another example of the document analysis apparatus 11 according to the embodiment of the present invention. The hardware configuration of the document analysis apparatus 11 is the same as that shown in FIG.

  The input processing unit 26 performs an input interface with the outside, and receives mail text data and commands through the communication control device 19 such as the Internet and the input device 20 such as the keyboard 23.

  The output processing unit 27 performs an output interface with the outside, and outputs mail body data through the communication control device 19 such as the Internet and the output device 17 such as the display device 18.

  The control unit 28 controls the entire apparatus. The control unit 28 stores the text data of the mail sent from the input processing unit 26 in the data temporary storage unit 29 of the storage device 25, the row combination determination threshold value calculation unit 34, and the combination. The line determination unit 35 and the data shaping unit 36 are controlled, and the contents of the data temporary storage unit 29 storing these calculation results are sent to the output processing unit 27 and output to the output device 17.

  The line combination determination threshold value calculation unit 34 reads the mail body data stored in the storage device 35 for each line, determines the data length for each line based on the number of characters in each line character string, and stores it in the storage device. At the same time, the difference between the maximum value of the data length stored in the storage device and the value of the data length having a high appearance frequency other than the maximum value is calculated as a row combination determination threshold value.

  The combined row determination unit 35 reads the body data of the mail stored in the storage device 25 for each row, and the data length of the read row is an approximate range of the maximum value. The data length of the row and the data of the next row When the difference from the length is the approximate range of the row combination determination threshold value calculated by the row combination determination threshold value calculation unit 34, the read line and the next line are given and stored with repetition information that is an index of row combination. It is stored in the device 25.

  The data shaping unit 36 reads the body data of the mail stored in the storage device 25 for each line, and when the combination between lines is instructed by the repetition information given by the combined line determination unit 35 of the read line, The combined character string is stored in the storage unit 25 by combining the lines.

  Next, the operation of another example of the document analysis apparatus according to the embodiment of the present invention shown in FIG. 7 will be described. FIG. 8 is a flowchart showing the operation of another example of the document analysis apparatus according to the embodiment of the present invention.

  Assume that the following mail body data A5 is input from the input processing unit 26.

“Mail text data A5”

  The mail body data A5 is obtained by cutting the following mail body data A6 at a position of 20 full-width characters according to the setting of the mail tool.

"Mail text data A6"

When mail body data A5 is input from the input device 20 to the arithmetic processing unit 12, the control unit 28 activates the input processing unit 26, and the input processing unit 26 stores the data in the storage unit 25 for each line feed of the mail body data A5. Data is read into the temporary data storage unit 29 (S801). Table 14 shows the mail body data A5 stored in the input data storage area of the temporary data storage unit 29.

  Next, the control unit 28 activates the row combination determination threshold value calculation unit 34. The line combination determination threshold value calculation unit 34 reads the mail body data A5 stored in the input data storage area of the data temporary storage unit 29 of the storage device 35 for each line, and for each line based on the number of characters in each line character string. The data length of each row is determined (S802), stored in the data temporary storage unit 29 of the storage device 35, and the appearance frequency is other than the maximum value of the data length stored in the data temporary storage unit 29 of the storage device 25 and the maximum value. The difference from the value of the high data length is calculated as the row combination determination threshold value (S803).

That is, the line combination determination threshold value calculation unit 34 reads the data for each line feed for the mail body data A5 in Table 14 and checks the data length of each line, and the result is mail body data as shown in Table 15. Is stored in the data temporary storage unit 29 of the storage device 25.

  Then, the data length maximum value and a data length having a high appearance frequency other than the maximum value are detected from the body data of the mail in Table 15, and a value obtained by subtracting the data length having a high appearance frequency from the maximum data length is calculated. It is set as a combination determination threshold value.

  Here, in the mail text data of Table 15, the maximum data length value is 40 and the data length with high appearance frequency is 10, so the row combination determination threshold value calculated in step S803 is 30.

  Next, the row combination determination unit 35 sets an initial value “1” to the row data designation variable i (S804), and sets an initial value “1” to a variable Q indicating repetition information (S805). The row data designation variable i is a variable i for acquiring the data Si of each row from the mail body data in Table 15, and the variable Q is a variable for determining row combination.

Then, the row combination determining unit 35 acquires the character string of the data Si and the length of the data Si from the mail body data in Table 15 (S806). As a result, the character string of the data Si (the setting of the mail tool is “2 at the time of mail transmission”) is acquired, and 40 is acquired as the length of the data Si. It is determined whether or not these data have been acquired ( S807) If the data cannot be acquired, “0” is set to the variable Q of the repetition information (S808), and the row combination determination unit 35 stores “0” in the repetition information of the mail body data in Table 16. The process ends.

  When the character string of the data Si and the length of the data Si can be acquired, the row combination determining unit 35 determines whether the length of the data Si is within the approximate range of the data length maximum value MAX (S809). The approximate range is obtained by subtracting the correction value from the data length maximum value MAX as the maximum approximate range determination value. MAX−correction value <| length of data Si | <MAX. Here, the correction value is set to 2, for example, but the user may be able to set an arbitrary value.

  When the length of the data Si is not within the approximate range of the data length maximum value MAX, the row combination determining unit 35 determines that the row of the data Si is not a repetition target and sets “0” to the variable Q of the repetition information. It is set (S810), 1 is added to the variable i to set i = i + 1 (S811), and the process returns to step S805. In this case, “0” is stored in the repetition information of the mail body data in Table 16.

  If it is determined in step S809 that the length of the data Si is within the approximate range of the data length maximum value MAX, the character string of the next data Si + 1 and the length of the data Si + 1 are acquired (S812). Thus, 10 is acquired as the character string of the data Si + 1 (0 character is self) and the length of the data Si + 1. If the data cannot be acquired, it is determined whether or not these data have been acquired (S813), the variable Q of the repetition information is set to “0” (S808), and the row combination determination unit 35 reads the text of the mail in Table 16 “0” is stored in the data repetition information, and the process ends.

  When the character string of the data Si + 1 and the length of the data Si + 1 can be acquired, the row combination determining unit 35 calculates the difference between the length of the data Si and the length of the data Si + 1 (S814). Subsequently, it is determined whether or not the difference is an approximate range and the row combination determination threshold value (S815). Here, since the difference is 30, it is the approximate value of the row combination determination threshold, so the value of variable Q (Q = 1) is set in the repetition information of data Si and data Si + 1 of the mail body data in Table 16. Then, Q = Q + 1 is set (S816).

  On the other hand, if it is determined in step S815 that the difference is not within the approximate range of the row combination determination threshold, “0” is set to the variable Q of the repetition information (S817), and the row combination determination unit 35 receives the mail of Table 16 “0” is stored in the repetition information of the body data. Then, it is determined whether or not the data Si + 1 is the last data (S818). If the data Si + 1 is the last data, the row combination determination unit 35 ends the process and proceeds to the next data shaping process (S819). . On the other hand, if it is not the last data, 2 is added to the variable i so that i = i + 2 (S820), and the process returns to step S806.

  Subsequently, a data shaping process is performed on the body data of the mail with the repetition information shown in Table 16 (S819). FIG. 9 is a flowchart showing the contents of the data shaping process of the data shaping unit 36. In the data shaping process in the data shaping unit 36, when combination between lines is instructed by the repetition information given to the read line, the line is combined to create a combined character string and the data is shaped. .

  First, “1” is set to the row data designation variable i for accessing the mail body data to which the repetition information shown in Table 16 is assigned, and “0” is set to the repetition state variable K (S901).

  Subsequently, the character string of the data Si and the repetitive information of the data Si are acquired from the mail body data shown in Table 16 (S902). It is determined whether or not these data can be acquired (S903). If the data cannot be acquired, the data shaping unit 36 ends the process. If the character string of the data Si and the repetition information of the data Si can be acquired, it is determined whether or not the repetition information is 0 (S904).

  If the repetitive information is 0, it is not combined with the next line. Therefore, the data Si is given a line feed and is shaped, and the data Si obtained as a result of the shaping is stored (S905). In this case, the data Si is stored in the shaping result storage area of the data temporary storage unit 29 of the storage device 25. Then, 1 is added to the variable i to set i = i + 1 (S906), and the process returns to step S902.

  If the repetition information is other than 0, the character string of the data Si + 1 next to the data Si and the repetition information of the data Si + 1 are acquired (S907). It is determined whether or not these data have been acquired (S908). If the data cannot be acquired, the data shaping unit 36 ends the process.

  When the character string of the data Si + 1 and the repetition information of the data Si + 1 can be acquired, the repetition information of the data Si + 2 next to the data Si + 1 is acquired (S909). It is determined whether or not the repetition information of the data Si + 2 has been acquired (S910). If the repetition information of the data Si + 2 has been acquired, it is determined whether or not it is 0 (S911).

  If the repetition information of Si + 2 cannot be acquired in the determination in step S910, or if the repetition information of data Si + 2 is 0 in the determination in step S911, it is determined whether or not the variable K is “1” (S912). If the value of the variable K is “1” in the determination in step S912, the character string of the data Si and the character string of the data Si + 1 are combined to add a line feed at the end, and the temporary data storage unit 29 of the storage device 25 Is stored in the shaping result storage area. That is, Si and Si + 1 are combined and stored (S913). When the repetition variable K is other than “1”, a line feed is added to the end of each of the data Si and the data Si + 1 and stored in the shaping result storage area of the data temporary storage unit 29 of the storage device 25. That is, Si and Si + 1 are stored (S914). Since the repeated state is canceled, “0” is reset to the variable K (S915).

  On the other hand, if it is determined in step S911 that the repetition information of the data Si + 2 is other than “0”, it can be determined that the data Si, the data Si + 1, the data Si + 2, and the data Si + 3 are consecutive lines according to a certain rule. Therefore, the character string of the data Si and the character string of the data Si + 1 are combined to give a line feed at the end, and stored in the shaping result storage area of the data temporary storage unit 29 of the storage device 25. That is, Si and Si + 1 are combined and stored (S916). Since it is in a repeated state, “1” is set to the variable K (S917). It is determined whether or not the data Si + 1 is the last data (S918). If the data is the last data, the process is terminated. If it is not the last data, 2 is added to the variable i to set i = i + 2 (S919). ), The process returns to S902.

When the processes in steps S901 to S919 are completed, mail body data as shown in Table 17 stored in the shaping result storage area of the data temporary storage unit 29 of the storage device 25 is obtained. As shown in Table 17, the character string is formatted.

  In this way, line breaks are automatically performed according to the settings of the mail tool, and even if the line breaks in the middle of the sentence and it becomes a state like a long sentence, short sentence, long sentence, short sentence, the data formatting of the sentence Sentences can be shaped by processing.

  According to the embodiment of the present invention, the sentence recognition processing is performed in the same hierarchy after shaping the hierarchy disturbed by the automatically inserted line feed, so that the sentence recognition accuracy can be improved.

The functional block diagram of the document analysis apparatus concerning embodiment of this invention. The block block diagram which shows the hardware constitutions of the document analysis apparatus which concerns on embodiment of this invention. The flowchart which shows operation | movement of Example 1 of the document analysis apparatus concerning embodiment of this invention. The flowchart which shows the processing content of the document hierarchy determination of the document hierarchy determination part in embodiment of this invention. The flowchart which shows the processing content of the sentence recognition process of the sentence recognition process part in embodiment of this invention. The flowchart which shows operation | movement of Example 2 of the document analysis apparatus concerning embodiment of this invention. The functional block diagram of another example of the document analysis apparatus concerning embodiment of this invention. The flowchart which shows operation | movement of another example of the document analysis apparatus of embodiment of this invention. The flowchart which shows the processing content of the data shaping of the data shaping part in embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 11 ... Document analysis device, 12 ... Arithmetic control device, 13 ... Processor, 14 ... Memory, 15 ... Document analysis program, 16 ... Work area, 17 ... Output device, 18 ... Display device, 19 ... Communication control device, 20 ... Input Device, 21 ... Mouse, 22 ... Keyboard, 23 ... Disk drive, 24 ... Hard disk drive, 25 ... Storage device, 26 ... Input processing unit, 27 ... Output processing unit, 28 ... Control unit, 29 ... Data temporary storage unit, 30 ... Document hierarchy determination unit, 31 ... Same hierarchy determination unit, 32 ... Document shaping unit, 33 ... Sentence recognition processing unit, 34 ... Line combination determination threshold value calculation unit, 35 ... Combined line determination unit, 36 ... Data shaping unit

Claims (6)

For each line character string that is read from the body data of the email input from the input device for each line feed and stored in the storage device , the data length is determined for each line based on the number of characters in each line character string, and each line character string is quoted A document hierarchy determination unit that determines the citation hierarchy for each line based on the number of symbols;
If the citation hierarchy of the data length approximation maximum line that is within the predetermined range of the maximum value among the data lengths determined by the document hierarchy determination unit is larger than the citation hierarchy of the next line, the data length approximation maximum line The length of the combined character string between the character string from which the quotation mark and line feed are removed and the character string from which the quotation mark and line feed are removed from the next line is the maximum value of the data length. It is determined whether or not it is within a predetermined range, and when the length of the combined character string is within a predetermined range of the maximum value of the data length, the actual hierarchy of the data length approximate maximum line and the next line is The same hierarchy determination unit for determining that the character strings are in the same hierarchy;
Of the line character strings stored in the storage device, the data length approximate maximum line determined to be the same hierarchy by the same hierarchy determination unit and the character string of the next line , the data length approximate maximum line A document analysis apparatus comprising: a document formatting unit that removes a quotation mark from the next line, shapes the line by approximating the end of the maximum data length approximate line, and outputs the entire line character string to the output device.   When the length of the combined character string is not within a predetermined range of the maximum value of the data length, the same layer determination unit determines the length of the reference layer of the data length approximate maximum line and the character string of the next line. And when the citation hierarchy of the data length approximate maximum line is larger than the length of the character string of the next line, the actual hierarchy of the data length approximate maximum line and the next line is a character string of the same hierarchy. The document analysis apparatus according to claim 1, wherein the document analysis apparatus determines that it is.   When the length of the combined character string is not within a predetermined range of the maximum value of the data length, the same layer determination unit determines a reference layer of the data length approximate maximum line and a predetermined reference layer threshold value. In comparison, when the citation hierarchy of the data length approximate maximum line is equal to or greater than a predetermined citation hierarchy threshold, the actual hierarchy of the data length approximate maximum line and the next line is a character string of the same hierarchy The document analysis apparatus according to claim 1, wherein the determination is performed.   When the combined character string length is not within a predetermined range of the maximum value of the data length, the same hierarchy determination unit determines the character string of the next line of the data length approximate maximum line and the predetermined number of characters. When the length of the character string in the next row of the data length approximate maximum row is equal to or less than the predetermined number of characters, the actual hierarchy of the data length approximate maximum row and the next row is the same The document analysis apparatus according to claim 1, wherein the document analysis apparatus determines that the character string is a hierarchical character string. A storage device that stores a document analysis program, an input device that inputs mail body data, an arithmetic control device that executes the document analysis program stored in the storage device, and outputs an arithmetic result of the arithmetic control device In a document analysis program used in a document analysis apparatus configured by a computer having an output device for
For each line string stored in the storage device, the body text data of the mail input from the input device is read for each line feed, and the data length is determined for each line based on the number of characters in each line string. A procedure for determining the citation hierarchy for each row based on the number of quotes,
If the citation hierarchy of the data length approximation maximum line whose data length is within the predetermined range of the determined data length is larger than the citation hierarchy of the next line, it is included in the character string of the data length approximation maximum line Whether the length of the combined character string between the character string from which the quotation mark and line feed are removed and the character string from which the quotation mark and line feed are removed is within the predetermined range of the maximum value of the data length A procedure for determining whether or not
When the length of the combined character string is within a predetermined range of the maximum value of the data length, it is determined that the actual hierarchy of the approximate maximum data length line and the next line is a character string in the same hierarchy. Procedure and
Among the line character strings stored in the storage device, for the character string of the data length approximate maximum line and the next line determined to be the same hierarchy, the citation of the line next to the data length approximate maximum line A document analysis program for executing a procedure of removing symbols and shaping by the process of combining at the end of the approximate maximum data length line and outputting the entire line character string to the output device. For each line character string that is read from the body data of the email input from the input device for each line feed and stored in the storage device , the data length is determined for each line based on the number of characters in each line character string, and each line character string is quoted Determine the citation hierarchy for each line based on the number of symbols,
If the citation hierarchy of the data length approximation maximum line whose data length is within the predetermined range of the determined data length is larger than the citation hierarchy of the next line, it is included in the character string of the data length approximation maximum line The length of the combined character string between the character string from which the quotation mark and the line feed are removed and the character string from which the quotation mark and the line feed are removed included in the next line is within a predetermined range of the maximum value of the data length. Whether or not
When the length of the combined character string is within a predetermined range of the maximum value of the data length, it is determined that the actual hierarchy of the data length approximate maximum line and the next line is a character string in the same hierarchy. ,
Among the line character strings stored in the storage device, for the character string of the data length approximate maximum line and the next line determined to be the same hierarchy, the citation of the line next to the data length approximate maximum line A document analysis method that removes symbols, shapes them by a process of combining them at the end of the approximate maximum data length line, and outputs the entire line character string to the output device.

Images