JP2003036261A - Device, program, and method for automatically classifying sentence, and computer readable recording medium recorded with automated sentence classifying program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an automated sentence classifying program and an automated sentence classifying device which can very rapidly and automatically classify sentences with high degree of accuracy, and further can classify all sentences into prescribed classification groups, which facilitate to grasp the subject of the sentences and other related items, and further to accumulate and transmit sentences by classification groups and to enable a user to smoothly utilize the sentences after the classification which enables to grasp the subjects and the like without decoding of the sentences. SOLUTION: This device analyzes sentences, extracts words, compares and checks a selected word against keyword groups which describe keywords by each classification items, adds up the appearance frequency matching with the extracted words by classification items, retrieves the classification items corresponding to the matched keywords, weights the keyword appearance frequencies by classification items by multiplying the coefficients set up by classification items, sorts the classification items with appearance frequencies modified with the relevant coefficients, adds up the sorted results, and calculates the classification items relevant to the sentence.

Description

Detailed Description of the Invention

[0001]

TECHNICAL FIELD The present invention relates to an automatic sentence classification program, a computer-readable recording medium having the automatic sentence classification program recorded therein, an automatic sentence classification method, and an automatic sentence classification device, and more particularly to extracting keywords included in a sentence. The present invention relates to a technique of totaling the appearance frequencies of the extracted keywords for each classification item and performing weighting by multiplying by a coefficient set for each classification item in advance to calculate the corresponding classification item and perform automatic classification. Furthermore, the present invention relates to a technique capable of calculating sub-information of a sentence and performing detailed classification.

[0002]

2. Description of the Related Art A conventional text automatic classification program, a computer-readable recording medium having the text automatic classification program recorded therein, a text automatic classification method, and a text automatic classification device use a feature pattern that characterizes a category to be classified. Those classified according to Japanese Patent Laid-Open No. 2000-22
2431), the frequency of keywords added to a sentence is automatically calculated, the keywords are associated with each other, a high-order directory is given to a keyword with a high frequency, and a keyword related to this keyword is created as a low-order directory to create a keyword tree structure. Then, a technique (Japanese Patent Laid-Open No. 2000-231560) for classifying each sentence according to a tree structure of a directory is proposed.

Further, a technique for classifying documents according to the degree of similarity to the semantic attributes of already classified sentences to which the semantic attributes of sentences are added using a corpus (Japanese Patent Laid-Open No. 2000-339310), and the sentences to be classified and attribute information are input. , A technique of extracting a keyword from a sentence, calculating a similarity between a classification rule composed of a combination of a keyword and attribute information and a keyword or attribute information for each classification group, and classifying into a group corresponding to the most similar classification rule (Patent application 2001-60199) A technique for checking whether or not one or more pre-registered keywords are included in an article, and saving an article including the keyword in a predetermined storage area provided for each keyword (Japanese Patent Laid-Open No. 2000-242242).
001-109772) and the like.

However, in these conventional examples, if the group to be classified is limited, the number of sentences that cannot be classified increases, and conversely, if all the sentences are tried to be classified, the number of classification groups becomes extremely large. There is a drawback that one sentence is classified into a large number of classification groups because it is classified according to the tree structure. Further, there is a drawback that the pattern, the keyword, or the sentence to be compared increases as the classification is repeated, and the classification process takes time.

Further, since only the appearance frequency or presence / absence of the keyword is directly reflected in the classification, the importance of each keyword to the classification group is not taken into consideration, or a certain keyword is excessive when complementing the keyword. However, there was a drawback that it was not possible to perform highly accurate classification.

Furthermore, it is difficult for the conventional technique to grasp the subject, related field, rough content, related country, region, company, etc. of the classified text from the classification group.
There was also a drawback that the text after classification could not be used smoothly. Further, when the classified documents are taken out from the storage place for each classification, that is, it is not possible to grasp what kind of classification the document belongs to by just looking at the document. .

[0007]

Therefore, according to the present invention, it is possible to automatically classify sentences very quickly and highly accurately, and further, to classify all sentences into predetermined classification groups, The post-classification group makes it easy to understand the subject, related fields, rough contents, related countries, regions, companies, etc. of the text, and the texts after the classification can be stored and transmitted for each classification. Therefore, users can grasp the subject etc. without deciphering the sentences and can use the sentences after classification smoothly. Furthermore, even if a computer terminal with a simple configuration is used, the above-mentioned effect can be obtained. An object is to provide an apparatus, a sentence automatic classification program, a computer-readable recording medium recording the sentence automatic classification program, and a sentence automatic classification method.

[0008]

As means for solving the above problems, firstly, the automatic sentence classification device of the present invention stores a classification item keyword table in which classification items and keywords associated with the classification items are described. Means, input means for inputting a sentence, text analyzing means for analyzing a sentence and extracting a word, and keyword scanning means for comparing and collating the extracted word with a keyword group of a classification item keyword table stored in a storing means An automatic sentence classification device characterized by having a comparison calculation means for weighting the result of comparison and collation and calculating a classification item to which the sentence corresponds.

Secondly, storage means for storing a classification item keyword table in which classification items and keywords associated with the classification items are stored, input means for inputting a sentence, and text analysis means for analyzing a sentence and extracting words. And a keyword scanning means for comparing and collating the extracted words with a keyword group in a classification item keyword table stored in the storage means, and summing up the frequency of appearance of keywords matching the extracted words for each classification item, and a weighting for the summing result. It is an automatic sentence classification device characterized by having a comparison operation means for adding a category and calculating a classification item to which a sentence corresponds.

Thirdly, the storage means stores a classification item coefficient table in which a coefficient is set for each classification item, and the comparison calculation means searches for a classification item corresponding to the matching keyword and classifies the keyword appearance frequency for each classification item. A sentence characterized by multiplying a coefficient set for each item and weighting, sorting the classification items by the appearance frequency corrected by the coefficient, and aggregating the classification results to calculate a classification item to which the sentence corresponds It is an automatic classification device.

Fourthly, a classification item keyword table that describes classification items consisting of at least two layers, large and small, and keywords associated with the small classification items, and classification item coefficients that set a coefficient for each large classification item for each attribute information. A storage means for storing a table; an input means for inputting a sentence having attribute information; a text analysis means for analyzing the inputted sentence to extract attribute information and a word from a portion other than the attribute information; The keyword scanning means for comparing and collating the selected words with the keyword group of the classification item keyword table stored in the storage means and summing up the appearance frequency of the keywords matching the extracted words for each small classification item, and the small classification corresponding to the keyword Search the major category item to which the item belongs, and set it for each major category item for each extracted attribute information to the keyword appearance frequency for each minor category item The weighting is performed by multiplying the coefficient of the large classification item to which the small classification item belongs, and the small classification items are sorted by the appearance frequency corrected by the coefficient, and the classification results are aggregated to classify the sentence. It is an automatic sentence classification device having a comparison calculation means for calculating an item.

Fifth, a classification item keyword table that describes classification items consisting of at least two layers, large and small, and keywords associated with the small classification items, and classification item coefficients that set a coefficient for each large classification item for each attribute information. A storage unit that stores a specific classification item keyword table in which a keyword designating a specific relationship between a table and classification items is described for each classification item, and an input unit that inputs a sentence having attribute information and heading information are input. The text analysis means for analyzing the sentence and extracting the attribute information and extracting the word from the part other than the attribute information is compared with the keyword group of the classification item keyword table stored in the storing means, and extracted. The frequency of appearance of keywords that match the specified word is aggregated for each subcategory, and the words extracted from the heading information are classified. Keyword scanning means for comparing and collating with the keyword group of the specific classification item keyword table stored in the means, and for searching the large classification item to which the small classification item corresponding to the keyword belongs, Of the coefficients set for each major classification item for each extracted attribute information, the keyword appearance frequency for each classification item is multiplied by the coefficient of the major classification item to which the minor classification item belongs The small classification items are sorted by frequency, the classification results are aggregated to calculate the classification items to which the sentence corresponds, and the comparison calculation means for calculating the small classification items corresponding to the keywords matching the words extracted from the heading information are provided. An automatic sentence classification device characterized by having.

According to the above first to fifth configurations, sentences can be classified very quickly and accurately with high precision, and further, all sentences can be classified into a predetermined classification group. The group after classification makes it easy to understand the subject matter and the rough content of the sentence.

Sixth, the storing means stores at least one or more code tables in which code items and keywords associated with the code items are stored, and the keyword scanning means stores at least one or more codes of the extracted words. By comparing and collating with the keyword group in the table, the appearance frequency of the keywords that match the extracted words is totaled for each code item, and the comparison operation means searches for the code item corresponding to the matching keyword, and the keyword appearance frequency for each code item It is an automatic sentence classification device characterized in that code items are sorted by and the code items corresponding to a sentence are calculated.

Seventh, the code table is at least one of a prefectural code table with prefectures as keywords, a country code table with country names as a keyword, and a company code table with company names as a keyword. It is an automatic sentence classification device characterized by including.

According to the sixth and seventh configurations, the group after the classification makes it easy to grasp the related fields, related countries, regions, companies and the like.

Eighth, the automatic document classifying apparatus further comprises output means for outputting the item of the classification result calculated by the comparison operation means.

Ninth, the automatic document classifying apparatus further comprises conversion result storage means for additionally writing the calculated classification result item in a sentence.

According to the above structure, the user can read the subject, related fields, rough contents, related countries, etc. without deciphering the text.
Areas, companies, etc. can be grasped and the sentences after classification can be used smoothly.

Tenth, the automatic document classification apparatus is an automatic text classification apparatus characterized in that it further comprises a storage means for storing the classified text for each item of the classification result and / or a distribution means for distributing the text.

According to the above configuration, the sentences after classification can be automatically stored and transmitted for each classification, and since the person does not perform the storage and the transmission, there is no error in the storage place and the destination.

Eleventh, the automatic document classifying apparatus further collects the classification results for a certain period for each item, statistically processes the appearance frequency of the set keyword, and normalizes by the number of target sentences for each classification item. The automatic sentence classification device characterized by having a classification result statistical means for performing statistical processing of the sentence distribution and an output means for outputting the statistical result.

Twelfth, the automatic document classifying apparatus further includes an output means for outputting the table stored in the storage means, an input means for reconstructing the table by changing, adding or deleting the contents of the table, and It is an automatic sentence classification device characterized by having storage means for storing a configured table.

According to the above configuration, the table used for classification can be reconfigured appropriately, so that sentences can be automatically classified with extremely high accuracy, and the group after classification can be
It becomes easy to understand the subject, related fields, rough contents, related countries, regions, companies, etc. of the sentence.

Thirteenth, an automatic sentence classification device composed of an editing terminal and a server system connected via a network, wherein the editing terminal has input means for inputting a sentence, the inputted sentence and a classification processing request. And an input means for receiving the classification processing result from the server system, and the server system stores a classification item keyword table in which keywords associated with the classification items are stored. An input means for receiving a sentence and a classification processing request from the editing terminal, a text analysis means for decomposing and extracting the sentence into words according to the classification processing request from the editing terminal, and the extracted words stored in the storing means. Keyword scanning means for comparing and collating with the selected keyword group, and the corresponding classification items by weighting the comparison results A sentence automatic classification apparatus characterized by comprising: a comparison operation means for finding, an output means for transmitting the classification processing result to the editing terminal.

According to the above method, it is possible to perform automatic sentence classification using a terminal having a simple structure connected to the network, and maintenance such as reconstructing a keyword group is required only once.

Fourteenth, a text analysis process of analyzing a sentence and extracting a word, a keyword scanning process of comparing and matching the extracted word with a keyword group described by associating a keyword with a classification item, and a result of the comparison and matching. It is an automatic sentence classification program characterized by causing a computer to execute a comparison calculation process for weighting and calculating a classification item to which a sentence corresponds.

Fifteenth, a text analysis process of analyzing a sentence and extracting a word, and comparing and collating the extracted word with a keyword group described by associating the keyword with a classification item, and appearance of a keyword matching the extracted word Automatic text writing characterized by causing a computer to execute a keyword scanning process for counting the frequency for each classification item and a comparison calculation process for weighting the totaled result to calculate and classify the classification item to which the sentence corresponds It is a classification program.

In the sixteenth comparison processing, the classification items corresponding to the matching keywords are searched, the keyword appearance frequency for each classification item is multiplied by a coefficient set for each classification item, and weighted. An automatic sentence classification program, characterized in that the classification items are sorted according to the corrected appearance frequency, and the classification results are totaled to calculate the classification item to which the sentence corresponds.

Seventeenth, a program for causing a computer to execute a process of automatically classifying a sentence having attribute information, analyzing a sentence having attribute information, extracting attribute information, and extracting a word from a portion other than the attribute information. It consists of two layers, a text analysis process to extract the extracted words and at least the small and large categories, and the keywords are compared and collated with the keywords described by associating the keywords with the small categories, and the keywords that match the extracted words appear. The keyword scanning process that aggregates the frequency for each small category item, and the large category item to which the small category item corresponding to the matching keyword belongs is searched, and the large category item for each attribute information extracted as the keyword appearance frequency for each small category item Of the coefficients set for each, the weighting is performed by multiplying by the coefficient of the large classification item to which the small classification item belongs, and the appearance frequency corrected by the coefficient To sort the items, a text automatic classification program, characterized in that sentence summarizes the classification results to perform the comparison operation process of calculating a classification item corresponding to the computer.

Eighteenth, a program for causing a computer to execute a process of automatically classifying sentences having attribute information and heading information, analyzing sentences having attribute information and heading information, extracting attribute information, and Text analysis processing that extracts words from parts other than information, and compares and collates the extracted words with the keyword group that describes the keywords by associating the keywords with the sub-category, and the appearance frequency of the keywords that match the extracted words The keyword scanning process that aggregates to, and the large classification items to which the small classification items corresponding to the matching keywords belong are searched, and the coefficient set for each large classification item for each attribute information extracted for the keyword appearance frequency for each small classification item Among them, the coefficient of the large classification item to which the small classification item belongs is multiplied and weighted, and the small classification item is sorted by the appearance frequency corrected by the coefficient. Then, the classification result is aggregated to calculate the classification item to which the sentence corresponds, and at the same time, the comparison calculation process of calculating the classification item to which the sentence corresponds by searching the classification item to which the keyword matching the word extracted from the heading information belongs It is an automatic sentence classification program characterized by being executed by a computer.

According to the above fourteenth to eighteenth means, it is possible to classify sentences extremely quickly and highly accurately and further, all sentences can be classified into a predetermined classification group. Then, by the group after the classification, it becomes easy to grasp the subject and the rough content of the sentence.

Nineteenth, in the keyword scanning process, the extracted words are compared and collated with at least one or more keyword groups in which the keywords are described by associating the keywords with the code items, and the frequency of appearance of the keywords matching the extracted words is determined by the code items. The comparison calculation process includes a process of searching for code items corresponding to the matched keywords, sorting the code items by the keyword appearance frequency of each code item, and calculating a code item to which the sentence corresponds. It is an automatic sentence classification program characterized by the following.

In the twentieth keyword group, which is described by associating keywords with code items, is a prefecture keyword group with prefectures, cities, towns and villages as keywords, country keyword groups with country names as keywords, and company keywords with company names as keywords. It is an automatic sentence classification program characterized by including at least one or more in a group.

According to the above nineteenth and twentieth means, the related fields, related countries, regions,
It becomes easy to grasp the companies.

In the twenty-first aspect, the automatic document classification program further includes an output process for outputting items of classification results calculated by the comparison calculation process, and causes the computer to execute these processes. It is an automatic sentence classification program.

In the twenty-second aspect, the automatic document classification program further includes a conversion result storing process for additionally writing the calculated result items of the classification into a sentence, and causes the computer to execute these processes. It is an automatic sentence classification program.

According to the above means, the user does not have to decipher the text, but the subject, related fields, rough contents, related countries,
Areas, companies, etc. can be grasped and the sentences after classification can be used smoothly.

Twenty-third, the automatic document classification program includes a process of accumulating and / or distributing the classified sentences for each item of the classification result, and causes the computer to execute these processes. It is an automatic sentence classification program.

According to the above means, it is possible to automatically store and transmit the classified sentences for each classification, and it becomes unnecessary for a person to perform the accumulation and the transmission, so that there is no error in the accumulation place and the destination.

Twenty-fourth, the document automatic classification program further collects the classification results for a certain period for each item, statistically processes the appearance frequency of the set keyword, and normalizes the classification result by the number of target sentences. An automatic sentence classification program characterized by including a classification result statistical process of performing a statistical process of the sentence distribution for each item and a process of outputting the statistical result, and causing a computer to execute these processes.

In the twenty-fifth aspect, a process for outputting a keyword group, a process for reconfiguring the keyword group by changing, adding, or deleting the keyword group and a process for storing the reconfigured keyword group are included. It is an automatic sentence classification program.

According to the above means, the table used for classification can be reconstructed as appropriate, so that sentences can be automatically classified with extremely high accuracy, and after the classification,
It becomes easy to understand the subject, related fields, rough contents, related countries, regions, companies, etc. of the sentence.

A computer-readable recording medium in which the automatic sentence classification program of the present invention is recorded in twenty-sixth is a computer-readable recording medium in which the automatic sentence classification program of the fourteenth to twenty-fifth means is recorded. Is.

In step 27, a step of inputting a sentence, a step of analyzing the sentence and extracting a word, and a step of comparing and collating the extracted word with a keyword group in which a keyword is described in association with a classification item, A method of automatically classifying a sentence, comprising the step of weighting the result of comparison and matching and finding a classification item to which the sentence corresponds.

Twenty-eighth, a text analysis step of analyzing a sentence and extracting a word, and comparing and matching the extracted word with a keyword group described by associating the keyword with a classification item and appearance of a keyword matching the extracted word A method for automatically classifying sentences, comprising: a keyword scanning step of totalizing the frequency for each classification item; and a comparison calculation step of weighting the totaled result to find a classification item to which the sentence corresponds.

In the twenty-ninth comparison operation step, a step of retrieving a classification item corresponding to the matching keyword and a step of multiplying the keyword appearance frequency of each classification item by a coefficient set for each classification item and weighting And a step of sorting the classification items by the appearance frequency corrected by the coefficient and totaling the classification results to calculate the classification item corresponding to the sentence.

A thirtieth method of automatically classifying sentences having attribute information, which is a text analyzing step of analyzing sentences having attribute information, extracting attribute information, and extracting words from a portion other than the attribute information. The extracted words consist of at least two layers of large and small classification items, and the keywords are compared and collated with the keyword group described in association with the small classification items, and the appearance frequency of the keywords matching the extracted words is calculated for each small classification item. The keyword scanning step for totaling and the large classification items to which the small classification items corresponding to the matching keywords belong are searched, and the coefficient set for each large classification item for each attribute information extracted for the keyword appearance frequency for each small classification item Among them, the coefficient of the large classification item to which the small classification item belongs is multiplied and weighted, and the small classification items are sorted by the appearance frequency corrected by the coefficient, and the classification result is displayed. A text automatic classification method characterized by having a comparison operation step of calculating a classification items sentences with meter corresponds.

The thirty-first method is a method for automatically classifying sentences having attribute information and heading information, in which a sentence having attribute information and heading information is analyzed, attribute information is extracted, and a portion other than the attribute information is extracted. A text analysis step that extracts words and a keyword scan that compares and collates the extracted words with a group of keywords that are described by associating the extracted keywords with subcategory items and totals the frequency of appearance of keywords that match the extracted words The step and the large classification item to which the small classification item corresponding to the matched keyword belongs are searched, and the small classification among the coefficients set for each large classification item for each attribute information extracted to the keyword appearance frequency for each small classification item Weighting is performed by multiplying by the coefficient of the large classification item to which the item belongs, the small classification items are sorted by the appearance frequency corrected by the coefficient, and the classification results are totaled and the sentence is summarized. Automatically calculating the corresponding classification item, and searching the classification item to which the keyword matching the word extracted from the heading information belongs, and calculating the classification item to which the sentence corresponds. It is a classification method.

According to the twenty-seventh to thirty-first configurations, it is possible to classify sentences very quickly and highly accurately and further, all sentences are classified into a predetermined classification group. It is possible to classify, and by the group after classification, it becomes easy to grasp the subject and rough content of the sentence.

In the thirty-second keyword scanning step, the extracted words are compared and collated with at least one or more keyword groups in which the keywords are described by associating the keywords with code items, and the appearance frequency of the keywords matching the extracted words is coded. Including the step of totaling for each item, the comparison operation step searches for the code item corresponding to the matched keyword, sorts the code items by the keyword appearance frequency for each code item,
A method for automatically classifying sentences, comprising a step of calculating a code item to which the sentence corresponds.

The keyword groups in which the keywords are described in association with the code items in the 33rd group are prefectures, cities, towns and villages keywords group, prefectures, cities, towns and villages keywords, country name keywords group, company name keywords companies. It is an automatic sentence classification method characterized by including at least one or more of a keyword group.

According to the thirty-second and thirty-third methods, it becomes easy to grasp the related fields, related countries, regions, companies, etc. by the group after classification.

In the thirty-fourth aspect, the automatic document classification method further comprises a classification result output step of outputting the classification result items calculated by the comparison operation step. .

The thirty-fifth aspect of the present invention is the automatic text classification method, further comprising a conversion result storage step of additionally writing the calculated classification result item in the text.

According to the above method, the user does not have to decipher the text, but the subject, related fields, rough contents, related countries,
Areas, companies, etc. can be grasped and the sentences after classification can be used smoothly.

In the thirty-sixth aspect, the automatic document classification method further comprises a step of accumulating and / or delivering the classified text for each item of the classification result. .

According to the above method, the sentences after the classification can be automatically stored and transmitted for each classification, and since the person does not perform the storage and the transmission, there is no error in the storage place and the transmission destination.

In the thirty-seventh aspect, in the above-mentioned automatic document classification method, the result of classification for a certain period is further tabulated for each item, the appearance frequency of the set keyword is statistically processed, and the classification result is normalized by the number of target sentences. An automatic sentence classification method characterized by including a classification result statistical step of performing a statistical process of the sentence distribution for each item and a step of outputting the statistical result.

In the thirty-eighth aspect, the steps of outputting a keyword group, reconfiguring the keyword group by changing, adding or deleting the keyword group, and storing the reconfigured keyword group are included. It is an automatic text classification method characterized by.

According to the above method, the table used for classification can be reconstructed appropriately, so that sentences can be automatically classified with extremely high accuracy, and the group after classification can be
It becomes easy to understand the subject, related fields, rough contents, related countries, regions, companies, etc. of the sentence.

[0062] In the thirty-ninth step, the step of inputting a sentence to the editing terminal, the step of transmitting the sentence to the server system to request the classification processing, and the step of the server system receiving the sentence and the classification processing request And a step in which the server system analyzes the sentence and extracts a word, a step in which the extracted word is compared and matched with a keyword group described by associating a keyword with a classification item, and the result of the comparison and comparison is weighted and applicable. The automatic text classification method is characterized by including a step of finding an item to be processed, a step of transmitting the classification processing result to the editor terminal, and a step of the editing terminal receiving the classification processing result.

According to the above method, it is possible to perform automatic sentence classification by using a terminal having a simple structure connected to the network, and maintenance such as reconstruction of the keyword group is required only once.

Here, the text is an electronic document, and the data format does not matter as long as it is a data format that can be processed by a computer. The sentences to be classified are not limited to Japanese and may be written in other languages such as English.

[0065]

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, an embodiment of an automatic document classification device of the present invention will be described in detail with reference to the drawings. FIG. 1 is a block diagram of an embodiment of the automatic sentence classification apparatus of the present invention. The automatic sentence classification apparatus 100 includes an input unit 110,
Text analysis means 120, keyword scanning means 130,
The comparison calculation unit 140, the output unit 150, the conversion result storage unit 160, the storage unit 170, the storage unit 180, and the control unit 190 are configured and can be configured by a personal computer or the like.

The input means 110 is composed of a keyboard, a scanner, a modem, a pointing device, etc., and is means for inputting sentences to be classified. Input is performed by typing a sentence with a keyboard, reading with a scanner, receiving from another terminal connected to the network via a modem, or the like. Further, it can also serve as means for changing, adding, deleting, and reconstructing various keyword groups displayed by the output means 150, instructing the storing means to store, and further instructing to change or cancel the classification result.

The text analysis means 120 is means for analyzing the input sentence, extracting attribute information and heading information, and extracting words from the sentence other than the attribute information part. When the sentence does not include the attribute information and the headline information, or when the sentence does not include the attribute information and the headline information, the sentence is analyzed without extracting the attribute information and the headline information, and the word is extracted. Sentence analysis is performed by reading the word analysis dictionary table 181 from the storage unit 18 and referring to this to perform morphological analysis or the like.

The attribute information is information for an extremely rough classification of the document inserted at an arbitrary place in the sentence, for example, the first line of the sentence. The headline information is the title of the document inserted at an arbitrary place in the text, for example, the second line of the text.

The keyword scanning means 130 is the storage means 18
This is a means for reading the classification item keyword table 182 stored in 0, comparing and collating with the words in the sentence extracted by the text analysis means 120, and totaling the appearance frequency of the keywords matching the extracted words for each small classification item. . Also, the keyword scanning means 130 associates the prefecture / city / town / village name serving as the keyword stored in the storage means 180 with the code number as the code item, and the prefecture / city / town / village code table 1
83, a country name code table 184 that associates a country name that is a keyword with a code symbol that is a code item, and a code table such as a company code table 185 that associates a company name that is a keyword and a code number that is a code item is read, and the text analysis means 120 It also serves as a means for comparing and collating with the words in the extracted sentence and totaling the appearance frequencies of the keywords that match the extracted words for each code item.

The comparison calculation means 140 searches the large classification items to which the small classification items corresponding to the matched keywords belong, and the coefficient set for each large classification item for each attribute information extracted as the keyword appearance frequency for each small classification item. Of these, the weighting is performed by multiplying by the coefficient of the large classification item to which the small classification item belongs, the small classification items are sorted by the appearance frequency corrected by the coefficient, the classification results are totaled, and the classification item to which the sentence corresponds or Further, it is means for calculating the code. With the sorting of minor categories,
The middle classification item and the large classification item to which the small classification item belongs are also sorted at the same time. Also, the keyword scanning means 1
The code items are sorted according to the frequency of appearance of the keywords for each code item aggregated in 30, and the code items are sorted by country information, prefecture information, company information, etc., and the classification results are aggregated, and the country information and the prefecture where the sentence corresponds It is also possible to calculate code items such as equal information and company information.

The output means 150 is from a monitor, a printer or the like for displaying or printing the classification items and / or the auxiliary information items calculated by the comparison operation means, or a modem or the like for delivering the text through the network. It is configured. The output means 150 is the classification item keyword table 1
82, various code tables 183, 184, 185 and coefficient table 186 can be output.

The conversion result storage means 160 is means for converting the classification result item into a classification code or adding it to the text as it is without conversion. The accumulating unit 170 is a unit that stores and accumulates the classified sentences in the storing unit 180 for each classification. The conversion result storage unit 160 and the storage unit 170 may not be provided. The storage unit 180 is composed of a magnetic recording medium such as a hard disk or a magneto-optical recording medium. The control unit 190 is a unit that controls the processing of each of the above units.

The storage unit 180 stores a word analysis dictionary table 181, a classification item as a keyword group in which keywords are described for each classification item, a classification item keyword table 182 in which codes indicating the classification items and keywords associated with the classification items are described. Prefectural municipality code table 18 in which a keyword group, which is a keyword group in which a keyword is described in association with a code item, is associated with a code number
3. A country name code table 184 that associates a country name with a code symbol, a company code table 185 that associates a company name with a code number, and a classification item coefficient table 186 are stored. Further, a specific information keyword table (not shown) in which keywords for specifying a specific relationship with the classification item, the country information item, the prefecture information, the company information item, and the company information item are described for each item It is also possible to store a specific classification item keyword table (not shown) in which the keywords designating the specific association of are described for each classification item. Also, various reconstructed tables 182, 1
83, 184, 185, 186 and the like are stored.

The storage unit 180 is the automatic sentence classification device 1.
00, not stored in 00, but stored in a server or the like having storage means, reception means, etc. connected to the automatic sentence classification device 100 via a network, and various tables 181, 182, 183, 184, 185 stored in the storage means. , 1
86 may be received and read by the automatic sentence classification device 100 when used, and may be transmitted to and stored in the server or the like when accumulating classified sentences. Further, when classifying only the classification items, the storage unit 180 stores the word analysis dictionary table 181 and the classification item keyword table 182.
And the classification item coefficient table 186 only, and the prefecture / city / town / village code table 183 and the country name code table 18 are stored.
4. The company code table 185 can be deleted as appropriate. In this case, only the classification items are searched, and information such as prefectures, towns, villages, country names, companies, etc. is not searched and classified. or,
The word analysis dictionary table 181 may not be provided depending on the sentence analysis and word extraction methods.

Further, the automatic sentence classification apparatus 100 collects the classification results for a certain period for each item, statistically processes the appearance frequency of the set keyword, and normalizes the number of target sentences to the sentence distribution of each item of the classified result. It is also possible to have a classification result statistic means 191 for performing the statistical processing of. The result of the statistics is output by the output means 150.

The classification item keyword table 182 is shown in FIG.
As shown in, there is a large item first, the large item is divided into medium items, and the medium item is also divided into small items to form a three-layer structure, and the keywords are associated with each small item. It has been described. The middle classification item is a superordinate concept of the small classification item, and the large classification item is a superordinate concept of the middle classification item and the small classification item. A corresponding code is attached to each item. It is not always necessary to have a three-layer structure, and one layer, two layers, or four layers or more may be appropriately used. As shown in FIG. 4, the prefecture / city / town / village code table 183 has a two-layer structure in which prefectures are in the upper layer and municipalities in each prefecture are in the lower layer, and code numbers are associated with each prefecture and each municipality. As shown in FIG. 5, the country name code table 184 is described by associating a country name with a code symbol. The company code table 185 is described in such a manner that companies and code numbers are associated with each other as shown in FIG.

In the classification item coefficient table 186, a large item is set for each attribute information, and a classification item coefficient is associated with each large item. The specific information keyword table is a table in which keywords that specify specific relationships with country name information, prefecture information, company information, and company information are described, and the keywords are associated with each item. The specific category item keyword table is a table in which keywords that specify a specific relationship with a category item are described, and a keyword is associated with each minor category item.

These various tables can be displayed on the monitor screen which is the output means 150, and can be changed, added, and deleted by the mouse, keyboard, etc., which are the input means 110, and various tables that have been changed, added, deleted, and reconfigured. Is stored in the storage unit 180 by the storage unit 170 according to an instruction from the input unit 110.

FIG. 2 is a block diagram of another embodiment of the automatic sentence classification apparatus of the present invention.
Is an editing terminal 21 composed of a personal computer or the like.
0, a server system 220 including an workstation and a personal computer, and an editing terminal 210
Network 23 connecting the server system 220 with the server system 220
It consists of zero. The editing terminal 210 is a keyboard,
An input unit 211 for inputting a sentence composed of a scanner, a modem, a pointing device, etc., and further for receiving classification information from the server system 220, a monitor,
It has at least output means 212 for displaying and printing classification results such as a printer and a modem, and further for transmitting sentences and classification requests to the server system 220, and control means for controlling these means. Other conversion result storage means and storage means may be included.

The server system 220 has a structure similar to that of the automatic sentence classification device 100, and includes a text analysis unit 221, a keyword scanning unit 222, and a comparison calculation unit 2.
23, input means 224 for receiving sentences and the like from an editing terminal composed of a modem and the like, editing terminal 210 for the classification result and the like.
To the output means 225, the word analysis dictionary table 241, the classification item keyword table 242, the prefecture code table 243, and the country code table 24.
4, a company code table 245 and a storage means 240 storing a classification item coefficient table 246, and a control means 226 for controlling these means. Other conversion result storage means 227, storage means 228, classification result statistical means 229, etc. may be provided. The network 230 does not matter whether it is wired or wireless.

Next, the processing procedure of the automatic document classification program of the present invention will be described. FIG. 8 is a flow chart showing an embodiment of the processing procedure of the automatic document classification program of the present invention.
First, a document to be classified is analyzed, and a text analysis process of extracting a word from the sentence is performed S10. Next, a keyword scanning process for comparing and collating the extracted words with the keyword group is performed S20. Then, a comparison calculation process of weighting the comparison result and calculating the item of the classification to which the sentence corresponds is performed S30. Next, a process of outputting the classification result after the classification item is found (S40). Next, it is necessary to judge whether the classification result needs to be changed or not. If it is necessary, change the document or prompt the change of the classification result. If no change is required or the classification result is changed. If so, the classification confirmation processing S50 is performed to proceed to the next step. Then, a conversion result storing process of additionally writing the classification result item in the sentence is performed in S60. Further, the process of accumulating and / or delivering the classified sentences for each classified item is performed S70.

The classification analysis output process S40 other than the text analysis process S10, the keyword scanning process S20, and the comparison calculation process S30, the classification confirmation process S50, the conversion result storage process S60, the accumulation process and / or the distribution process S70 are not always necessary. However, these processes may be omitted, and an appropriate process may be selected from these processes to proceed, or the process may be proceeded by appropriately changing the order.

To explain each process in detail, in S10 which performs the text analysis process, a word analysis dictionary for decomposing sentences to be classified into words is read S101, attribute information is extracted from the sentences S102, finding information from the sentences is read. S103, the word in the found information part is extracted by referring to the word analysis dictionary, and the word in the body part from which the attribute information and the found information are removed from the sentence is extracted by referring to the word analysis dictionary in S104. . Note that attribute information extraction S102, finding information word extraction S103, and body text word extraction S1.
04 does not necessarily have to be in this order, and may be processed simultaneously. If the text to be classified does not have attribute information or finding information, or if it is not reflected in the classification, the processing of attribute information extraction S102 and finding information word extraction S103 is appropriately omitted. Note that the attribute information, the finding information, and the body portion are divided by control codes such as tags and other symbols, and delimiter information such as line feeds, and each portion is extracted.

The extraction of words in the text analysis process S10 uses various other extraction methods in addition to analyzing the sentence by extracting a word based on the morphological analysis with reference to the word analysis dictionary as described above. However, the extraction method is not limited.

In step S20 for performing the keyword scanning process, in addition to the classification item keyword table or the classification item keyword table forming the keyword group, an item table of attached information such as a prefecture code, country code table, company code table, etc. Of these, at least one is read S201, a suitable code table is expanded in the memory in addition to the classification item keyword table or the classification item keyword table, and the extracted word is extracted into the classification item keyword table or the keyword group of the code table. S203, and the appearance frequency of the keyword of the keyword group of the classification item keyword table that matches the extracted word is aggregated for each small classification item, or the appearance frequency of the keyword of the keyword group of the code table is further calculated for each code item. S2 to total 4. Even if there are a plurality of the same keywords, the number is counted. The keywords included in the headline information may or may not be counted.

S30 for carrying out the comparison calculation process is each classification item of large, medium, and small corresponding to the keyword that matches the extracted word, or in addition, a code item of country name information, a code item of prefecture information, a code item of company information. Search for S
301, the keyword appearance frequency for each searched small classification item is multiplied by the coefficient of the large classification item to which the small classification item belongs among the coefficients set for each large classification item for each extracted attribute information, or Further, the keywords for each code item searched for each code table are weighted by the appearance frequency and S
302, the small classification item is sorted by the appearance frequency corrected by the coefficient, and as a result, the large and medium classification items to which the small classification item belongs are sorted, or the code items are further sorted by the appearance frequency of each code item, and S303, the classification item, Country name information code items, prefecture / city / town / village information code items, company information code items are aggregated into classification items or their code items, country / name information code items, prefecture / town / town / village information code items, companies S30 of calculating information code item
6.

Of course, only classification items may be searched, sorted, etc., and processing such as country code information code items, prefecture / city / town / village information code items, and company information code items may be omitted as appropriate. It may be that.
Further, the weighting S302 is based on the keyword appearance frequency for each searched small classification item, the medium classification item or the small classification item to which the small classification item belongs among the coefficients set for each extracted attribute information or the small classification item. Weighting may be performed by multiplying by the coefficient of the classification item. Further, the weighting does not reflect attribute information with or without attribute information, and the weighting is not reflected for each searched classification item and for each small classification item. The keyword appearance frequency may be weighted by multiplying it by a coefficient set for each large classification item to which the small classification item belongs, or for each small classification item, or for each intermediate classification.

If the extracted word matches the keyword of the keyword group that specifies a specific relationship with at least one of the classification item, the country name information, the prefecture information, the company information, and the company information in advance, A classification item for which a specific association is designated, a code item for country name information, a code item for prefecture information, and a code item for company information are searched by referring to the specific information keyword table, and S304,
It may be reflected in the classification result. In this case, it is possible to arbitrarily set which of the classification item, country name information, prefecture / town / village information, and company information retrieved in S303 to be reflected in the classification result. Furthermore, when the found information is subjected to word extraction processing separately from the text,
If the keywords in the keyword group of the specific category item keyword table that specify a specific relationship with the category item in advance match the words extracted from the finding information, specify the particular category item from the keywords in the finding information part. It may be S305 to search. In this case, it is possible to arbitrarily set which of the classification results sorted in S303 and prioritized to be reflected in the classification result.

It should be noted that, as for the classification result, only the sorted highest item may be calculated as the classification result.
A plurality of appropriate upper items may be calculated as the classification result.

Processing for outputting the result of classification S40
Reads the classification item keyword table or the various code tables in addition to S401, searches the classification item keyword table or the various code tables for the item name from the classification result code in S402, and displays in the classification result display dialog. The item name is set in S403, and the classification result is displayed or further printed in S404. Here, as for the classification result, only the sorted highest item may be displayed as the result, but a plurality of appropriate upper items may be displayed as the classification result.

In the classification confirmation processing S50, when the classification result is displayed, it is determined whether or not the classification result needs to be changed and whether or not the classification result is cancelled. S501, if no change is necessary or the classification result is changed. When the change is received in response to the instruction of S502, the process proceeds to the next process, and when the classification result is cancelled, the document is prompted to be changed S503 and the classification confirmation process S50 is performed.

In step S60, the conversion result storing process is performed by converting the item name of the classification result into a code on the table by referring to the classification item keyword table or various code tables in addition, and in step S601, a text format sentence such as XML data. A code converted from the classification result is set in a predetermined tag item of No. and added to the sentence S602. Of course, the word of the item of the classification result may be added as it is without converting the classification result into a code.

In step S70, the process of accumulating and / or delivering the sentences for each classification is performed.
S701 which accumulates ML data and / or delivers the sentence to which the classification result was added S702. The accumulation is stored in advance in a storage area for each classification result, but may be stored without providing a storage area for each classification result. Further, although the distribution is distributed to the distribution destination corresponding to the predetermined classification, it may be distributed to any distribution destination regardless of the classification. The classification results may be any of large classification items, middle classification items, small classification items, country name information, prefecture information, company information, and company information, or any of these classifications. The storage or distribution process may be performed in correspondence with the item of or all applicable categories.

Further, a classification result totaling process of totaling the results of classification for a certain period of time, statistically processing the appearance frequency of the set keyword, and statistically processing the sentence distribution of each item of the classification result normalized by the number of target sentences. It is recommended to include S801 and processing S802 for outputting the totalized result.

Further, a classification item keyword table which is a keyword group described for each classification item or / and various code tables which are a keyword group described for each code item are output and S901, the keyword group is changed, added or deleted. It is recommended to include S902, S903 for reconstructing the keyword group, and processing.

Further, as the computer-readable recording medium in which the automatic document classification program is recorded, a computer-readable recording medium such as a magnetic disk, a magnetic tape, or an optical disk is used for the automatic sentence classification program for causing a computer to execute the above-described processing. It is recorded on a medium. When performing the automatic sentence classification process, the recording medium is read by a computer and the program is executed to perform the automatic sentence classification process as described above.

Next, one embodiment of the automatic classification method of the present invention will be specifically described by taking the sentence 900 shown in FIG. 18 as an example. First, the text analysis means 120 reads the word analysis dictionary table 181 for analyzing the sentence input by the input means 110 and extracting words from the storage means 180, S1.
The attribute information 910, the headline information 920, and the body portion of the sentence are recognized by the 01, the control code such as the tag, other symbols, and the delimiter information such as a line feed, and the attribute information “foreign communication” 910.
S102, the word "trading" of the found information 920 is extracted by referring to the word analysis dictionary table 181, and S10 is extracted.
3, referring to the word analysis dictionary table 181, the words "parliament", "financial", "committee", "government" in the text,
"Trade", "Trade", "Trade", "Public hearing", "US", "Osaka", "Suzuki industry", etc. are extracted in S104.
(Text analysis process S10). Incidentally, extraction S102 of the attribute information 910, extraction S103 of the word of the finding information 920,
The extraction of the words in the body portion S104 does not necessarily have to be performed in this order, and they may be processed simultaneously. or,
If the text to be classified does not have the attribute information 910 or the finding information 920, the processing of extracting attribute information S102 and finding information word S103 is appropriately omitted.

Next, the keyword scanning means 130, in addition to the classification item keyword table 182 or the classification item keyword table 182, the prefecture code code table 1
83, country name code table 184, company code table 185, and the like, at least one of the code tables is read in S201, the classification item keyword table 182 or the classification item keyword table 182, and an appropriate code table is expanded in the memory. S202, the extracted words "parliament", "finance", "committee", "government", "trade",
"Public hearing", "US", "Osaka", "Suzuki industry", etc. are the keywords of the category item keyword table 182 "Administration", "Finance", "Trade", etc., and also the keywords of the prefecture code table 183. Compares "Hokkaido", "Sapporo", "Osaka", etc. with the keywords "Afghanistan", "US" etc. of the country code table 184 and the keywords "Suzuki Sangyo", "Tanaka Bank" etc. of the company code table 185. S203, the keywords “trade”, “trade”, “import”, “steel”, which match the extracted words,
S204 (keyword scanning process S20) in which the appearance frequencies of "US", "Osaka", "Suzuki industry", etc. are totaled for each classification item, country code item, prefecture code, company code item. In this embodiment, as shown in FIG. 19, "trade" is 3, "trade" is 2, "import" is 1, and steel is "steel".
2, “automobile” is 2, “government” is 2, “finance” is 1,
"Red" is 1, "US" is 1, "Osaka" is 1,
"Suzuki industry" is 1.

Then, the comparison operation means 140 refers to various tables and matches the keyword "trade" with the large classification item "economy", the middle classification item "macro economy", and the small classification item "trade". The classification item corresponding to the keyword, the country name information, the prefecture information, the company information is searched and S301 is searched, and the keyword appearance frequency for each small classification item, for example, "trade" in the small item "trade" Since "3", "Trade" is 2, and "Import" is 1, the coefficient set in the major category "Economy" to which the minor category "Trade" of the attribute information "Foreign Communication" 910 extracted in "6" belongs The weighting is performed by multiplying by "0.10", and the same processing is performed for each subcategory item, and the appearance frequency of the country code item, the prefecture code, the company code item, for example, "US" is 1 and "Osaka""But , “Suzuki Sangyo” totals 1 and sorts the classification items by S302, the appearance frequency “0.6” corrected by the coefficient, and the appearance frequency of country name information, prefecture information, company information for each information. Sorted into S
303, a classification item, country name information, prefecture / city / town / village information, and company information are totaled to calculate a classification item, country name information, prefecture / town / town / village information, and company information S306 (comparison calculation processing S30). At this time, various calculated items include, for example, “economy” is “02000”, “US” is “USA”,
"Suzuki industry" is converted into a code like "2501".

It should be noted that the classification information, the country name information, the prefecture / town / town / village information, and the company information in which the keyword and the specific association are designated in advance are searched with reference to the specific information keyword table stored in the storage unit 180, S304, It may be reflected in the classification result of the classification item, country name information, prefecture information, company information. In this case S301
It is possible to arbitrarily set which of the country name information, the prefecture / town / town / village information, and the company information retrieved in step 1 to be reflected in the classification result. Further, when the found information is subjected to the word extraction processing separately from the text, the specific classified item keyword table stored in the storage means 180 is referred to search for a specific classified item from the keyword of the found information. S to do
It may be 305. In this case, it is possible to arbitrarily set which of the classification results sorted in S303 and prioritized to be reflected in the classification result.

Next, the output means 150 outputs the classification item keyword table 182 or the various code tables 1
82, 183, 184, 185 is read in S401, the classification item keyword table or the various code tables 182, 183, 184, 185 is referred to, and the classification result code information "02000" to "economy" is changed to "US".
Search for "A" from "A", "2501" to "Suzuki Sangyo", etc. S402, and set the item names "Economy", "US", "Osaka", "Suzuki Sangyo", etc. in the classification result display dialog. S403, S404 for displaying or printing the classification result as shown in FIG. 20 (classification result output process S40).

Next, when the output means 150 displays the classification result, for example, the input means 110 displays the display screen 15
The designation of the OK tag 158 or the cancellation tag 159 or the reference tag 157 on 1 indicates whether or not the change of the classification result is necessary, S501, and if the change is not necessary, the OK tag 158 is specified and the next process is performed. Prompt and reference tag 1
57 and the item list (not shown) displayed on the screen 151 according to the designation, the change process is performed. If the cancel cell tag 159 is designated, the text is changed. S503 requesting a change (classification confirmation processing S50).

Then, the conversion result storage means 160 refers to the classification item keyword table 182 or the various code tables 182, 183, 184 and 185, and the item names "economy", "US" and "Suzuki Sangyo" of the classification result. , Etc. from the code "02000", "USA", "250"
1 ”or the like in step S601, and the code converted from the classification result is set in a predetermined tag item of a text-formatted sentence in step S601.
602 (conversion result storage processing S60). It should be noted that the item names “economy”, “US”, “Suzuki Sangyo”, etc. of the classification result may be added to the text as they are without being converted into codes.

Then, the storage means 170 stores the sentence 999 to which the classification result as shown in FIG. 21 is added.
S701 and / or output means 1 such as a modem
Reference numeral 50 denotes a sentence 999 stored in the storage unit 180 or to which a classification result not yet stored is added, which is distributed via the network S702.

The classification result statistical means 191 totals the classification results for a certain period for each item, statistically processes the appearance frequency of the set keyword, and statistically processes the sentence distribution for each classified item normalized by the number of target sentences. S801, output means 150
Outputs the statistical result in S802.

Incidentally, various tables 181, 182, 183, 184, 18 are dealt with in response to changes in words, changes in area names, establishment of companies, mergers, disappearances, etc., with the changes of the times.
5, 186 are read from the storage unit 180 according to an instruction from the input unit 110 and displayed on the output unit 150 in S901, and items, keywords, coefficients, etc. are added, deleted, or changed in the input unit 110, and S902 is reconstructed. S90
3. The reconstructed table is stored in the storage unit 180.

[0107]

As described above, according to the present invention, it becomes possible to classify sentences extremely quickly and highly accurately. Also, all sentences can be classified into a predetermined classification group, and further, after the classification,
It is easy to understand the subject, related fields, rough contents, related countries, regions, companies, etc. of the text, and the texts after the classification can be stored and transmitted for each classification, and the user can also write the texts. It becomes possible to smoothly utilize the classified texts that can grasp the subject without deciphering the above, and further, the above-mentioned effects can be obtained even by using a computer terminal having a simple configuration.

[Brief description of drawings]

FIG. 1 is a block diagram of an embodiment of an automatic sentence classification device of the present invention.

FIG. 2 is a block diagram of another embodiment of the automatic sentence classification device of the present invention.

FIG. 3 is a diagram showing a classification item keyword table according to an embodiment of the present invention.

FIG. 4 is a diagram showing a prefecture / city / town / village code table according to an embodiment of the present invention.

FIG. 5 is a diagram showing a country code table according to an embodiment of the present invention.

FIG. 6 is a diagram showing a company code table according to an embodiment of the present invention.

FIG. 7 is a diagram showing a coefficient table according to an embodiment of the present invention.

FIG. 8 is a flow chart showing the flow of processing according to an embodiment of the present invention.

FIG. 9 is a flowchart showing a part of a processing flow of the embodiment of the present invention.

FIG. 10 is a flowchart showing a part of a processing flow of the embodiment of the present invention.

FIG. 11 is a flowchart showing a part of the processing flow of the embodiment of the present invention.

FIG. 12 is a flowchart showing a part of the processing flow of the embodiment of the present invention.

FIG. 13 is a flowchart showing a part of a processing flow of the embodiment of the present invention.

FIG. 14 is a flowchart showing a part of the processing flow of the embodiment of the present invention.

FIG. 15 is a flowchart showing a part of the processing flow of the embodiment of the present invention.

FIG. 16 is a flowchart showing a part of the processing flow of the embodiment of the present invention.

FIG. 17 is a flowchart showing a part of the processing flow of the embodiment of the present invention.

FIG. 18 is a diagram showing an example of a classification target sentence in the present invention.

FIG. 19 is a conceptual diagram of weighting of classification items according to an embodiment of the present invention.

FIG. 20 is a diagram showing a display screen of classification results according to the embodiment of the present invention.

FIG. 21 is a diagram showing classified sentences according to an embodiment of the present invention.

[Explanation of symbols]

100 Automatic sentence classifier 110 input means 120 Text analysis means 130 keyword scanning means 140 Comparison calculation means 150 output means 160 conversion result storage means 170 storage means 180 storage means 182 Classification item keyword table 183 Prefecture Code Table 184 country code table 185 Company Code Table 186 Classification item coefficient table 190 control means 191 Classification result statistical means 200 Automatic sentence classifier 210 Editing terminal 211 Input means 212 Output means 220 server system 230 network 910 Attribute information 920 Headline information

Claims (39)

[Claims] 1. A storage unit for storing a classification item keyword table in which classification items and keywords associated with the classification items are stored, an input unit for inputting a sentence, and a text analysis unit for analyzing a sentence and extracting a word. A keyword scanning means for comparing and collating the extracted words with a keyword group of a classification item keyword table stored in the storing means, and a comparison calculating means for weighting the result of the comparison and collation to calculate a classification item to which the sentence corresponds. An automatic sentence classification device characterized by having. 2. A storage unit for storing a classification item keyword table in which classification items and keywords associated with the classification items are stored, an input unit for inputting a sentence, and a text analysis unit for analyzing a sentence and extracting a word. , A keyword scanning means for comparing and collating the extracted words with a keyword group of a classification item keyword table stored in the storage means, and summing up the appearance frequency of keywords matching the extracted words for each classification item, and weighting the summing result An automatic sentence classification device, comprising: a comparison calculation means for calculating a classification item to which a sentence corresponds. 3. The storage means stores a classification item coefficient table in which a coefficient is set for each classification item, and the comparison calculation means searches for a classification item corresponding to a matching keyword, and the classification item is classified according to the keyword appearance frequency for each classification item. The weighting is performed by multiplying by a coefficient set for each, the classification items are sorted by the appearance frequency corrected by the coefficient, the classification results are totaled, and the classification item to which the sentence corresponds is calculated. The automatic sentence classification device according to claim 1 or claim 2. 4. A classification item keyword table in which a classification item having at least two layers, large and small, and a keyword associated with the small classification item are described, and a classification item coefficient table in which a coefficient is set for each large classification item for each attribute information. Storage means for storing, and input means for inputting a sentence having attribute information,
A text analysis means that analyzes the input sentence, extracts attribute information and extracts words from the part other than the attribute information, and compares and collates the extracted words with the keyword group of the classification item keyword table stored in the storage means. , A keyword scanning means for totaling the appearance frequencies of keywords matching the extracted words for each small classification item, and searching for a large classification item to which the small classification item corresponding to the keyword belongs, and extracting the keyword appearance frequency for each small classification item Of the coefficients set for each large classification item for each attribute information, the weighting is performed by multiplying the coefficient of the large classification item to which the small classification item belongs, and the small classification items are sorted by the appearance frequency corrected by the coefficient, An automatic sentence classification device, comprising: a comparison calculation unit that aggregates classification results and calculates a classification item to which a sentence corresponds. 5. A classification item keyword table in which a classification item consisting of at least two layers, large and small, and a keyword associated with the small classification item is described, and a classification item coefficient table in which a coefficient is set for each large classification item for each attribute information. Storage means for storing a specific classification item keyword table in which a keyword for designating a specific relation between a classification item and a classification item is stored, and input means for inputting a sentence having attribute information and heading information. A text analysis means that analyzes a sentence and extracts attribute information and extracts words from a portion other than the attribute information is compared with the keyword group of the classification item keyword table stored in the storage means and extracted. The frequency of appearance of keywords that match the words is totaled for each small classification item, and the words extracted from the heading information are stored. And keyword scanning means and keywords of stored specified category keyword table comparison matches, searching for keywords that match the extracted words,
A large classification item to which the small classification item belongs among the coefficients set for each large classification item for each attribute information extracted for the keyword appearance frequency for each small classification item by searching for the large classification item to which the small classification item corresponding to the keyword belongs The weighting is performed by multiplying the coefficient of the item, the small classification items are sorted by the appearance frequency corrected by the coefficient, the classification results are aggregated to calculate the classification item to which the sentence corresponds, and the word extracted from the heading information is used. An automatic sentence classification device comprising: a comparison calculation means for calculating a small classification item corresponding to a matching keyword. 6. The storage means stores at least one or more code tables describing code items and keywords associated with the code items, and the keyword scanning means stores at least one or more code tables of extracted words. Comparing and collating with the keyword group of, the appearance frequency of the keyword that matches the extracted word is totaled for each code item, and the comparison operation means searches for the code item corresponding to the matching keyword, and the keyword appearance frequency for each code item is used. 6. The automatic sentence classification device according to claim 1, wherein the code items are sorted and the code item corresponding to the sentence is calculated. [Claim 7] The code table includes at least one of a prefectural municipality code table having a prefecture name as a keyword, a country code table having a country name as a keyword, and a company code table having a company name as a keyword. The automatic sentence classification device according to claim 6, characterized in that. 8. The automatic document classification device according to claim 1, wherein the automatic document classification device further comprises output means for outputting items of classification results calculated by the comparison operation means. . 9. The automatic document classification apparatus according to claim 1, wherein the automatic document classification apparatus further includes conversion result storage means for additionally writing the calculated classification result item in a sentence. 10. The automatic document classification device further comprises a storage means for storing the classified texts and / or a distribution means for distributing the classified texts for each item as a result of the classification.
10. The automatic document classification device according to claim 9. 11. The document automatic classification apparatus further collects the classification results for a certain period for each item, statistically processes the appearance frequency of the set keyword, and normalizes the number of target sentences by the sentence for each classification item. 11. The automatic sentence classification device according to claim 1, further comprising a classification result statistical means for performing statistical processing of distribution and an output means for outputting the statistical result. 12. The automatic document classification device is further reconfigured with an output means for outputting a table stored in a storage means, and an input means for reconfiguring the table by changing, adding or deleting the contents of the table. 12. The automatic document classification device according to claim 1, further comprising storage means for storing the table. 13. An automatic text classification device comprising an editing terminal and a server system connected via a network, wherein the editing terminal is an input means for inputting a text and a server for inputting the input text and a classification processing request. The server system has an output means for transmitting to the system and an input means for receiving the classification processing result from the server system, and the server system has a storage means for storing a classification item keyword table describing keywords associated with the classification items, and an editing An input means for receiving a text and a classification processing request from the terminal,
A text analysis means for decomposing and extracting the sentence into words according to a classification processing request from the editing terminal, a keyword scanning means for comparing and collating the extracted words with a keyword group stored in the storing means, and weighting the comparison result. An automatic sentence classification device comprising: a comparison calculation means for adding a corresponding classification item to find out a corresponding classification item; and an output means for transmitting a classification processing result to an editing terminal. 14. A text analysis process of analyzing a sentence and extracting a word, a keyword scanning process of comparing and matching the extracted word with a keyword group described by associating a keyword with a classification item, and weighting the result of the comparison and matching. An automatic sentence classification program characterized by causing a computer to execute a comparison calculation process for calculating a classification item to which a sentence corresponds. 15. A text analysis process of analyzing a sentence and extracting a word, and comparing and collating the extracted word with a keyword group described by associating the extracted word with a classification item to determine the appearance frequency of the keyword that matches the extracted word. An automatic sentence classification program characterized by causing a computer to execute a keyword scanning process for totalizing each classification item and a comparison calculation process for weighting the totaled results to calculate and classify classification items corresponding to sentences . 16. The comparison operation process searches for a category item corresponding to a matching keyword, multiplies the keyword appearance frequency of each category item by a coefficient set for each category item, and weights it, and corrects with the coefficient. Sort the categories by appearance frequency,
16. The process of totaling classification results to calculate a classification item to which a sentence corresponds, 16.
The automatic sentence classification program described in. 17. A program for causing a computer to automatically classify sentences having attribute information, analyzing sentences having attribute information, extracting attribute information, and extracting words from a portion other than the attribute information. It consists of a text analysis process, and the extracted words are composed of at least two layers of large and small classification items.The keywords are compared and collated with the keyword group described in association with the small classification items, and the appearance frequency of the keywords that match the extracted words is determined. A keyword scanning process that aggregates each small classification item and a large classification item to which the small classification item corresponding to the matching keyword belongs is searched, and the keyword appearance frequency for each small classification item is extracted for each large classification item for each attribute information extracted Of the set coefficients, weighting is performed by multiplying by the coefficient of the large classification item to which the small classification item belongs, and the small classification item with the appearance frequency corrected by the coefficient. Sorted, text automatic classification program sentences by aggregating the classification results, characterized in that to perform the comparison operation process of calculating a classification item corresponding to the computer. 18. A program for causing a computer to execute a process of automatically classifying sentences having attribute information and heading information, analyzing a sentence having attribute information and heading information, extracting attribute information, and excluding attribute information. A text analysis process that extracts words from the part of the word is compared with the keyword group that describes the extracted words by associating the keywords with the subcategory, and the appearance frequency of the keywords that match the extracted words is aggregated for each subcategory. Of the coefficients set for each large classification item for each attribute information extracted to the keyword scanning process and for searching the large classification items to which the small classification items corresponding to the matching keywords belong, and the keyword appearance frequency for each small classification item The coefficient of the large classification item to which the small classification item belongs is multiplied and weighted, and the small classification item is sorted by the appearance frequency corrected by the coefficient, A comparison calculation process is performed to calculate the classification items to which the text corresponds by calculating the classification items to which the sentence corresponds and to calculate the classification items to which the sentence corresponds by searching for the classification item to which the keyword matching the word extracted from the heading information belongs. An automatic text classification program characterized by being executed. 19. The keyword scanning process compares and compares the extracted words with at least one or more keyword groups in which the keywords are described by associating the keywords with the code items, and the appearance frequency of the keywords matching the extracted words is determined for each code item. The comparison calculation process includes a process of searching for code items corresponding to the matched keywords, sorting the code items by the keyword appearance frequency for each code item, and calculating a code item to which the sentence corresponds. 15. The method according to claim 14,
A sentence automatic classification program according to claim 18. 20. A keyword group in which a keyword is described in association with a code item is a prefecture keyword group that uses a prefecture name as a keyword, a country keyword group that uses a country name as a keyword, and a company keyword group that uses a company name as a keyword. 20. The automatic sentence classification program according to claim 19, wherein at least one of them is included. 21. The automatic document classification program further includes an output process for outputting the classification result items calculated by the comparison operation process, and causes the computer to execute these processes. The automatic sentence classification program described in 14 to 20. 22. The automatic document classification program further includes a conversion result storage process of additionally writing the calculated classification result item in a sentence, and causes a computer to execute these processes. 21. The automatic sentence classification program described in any one of 21 to 21. 23. The automatic document classification program includes a process of accumulating and / or distributing the classified sentences for each item of the classification result, and causes a computer to execute these processes. The automatic sentence classification program according to any one of claims 14 to 22. 24. The automatic document classification program further collects the classification results for a certain period for each item, statistically processes the appearance frequency of the set keyword, and normalizes the classification result by the number of target sentences for each item. 24. The automatic text classification according to claim 14, further comprising: a classification result statistical processing for performing statistical processing of the sentence distribution and a processing for outputting the statistical result, the processing being executed by a computer. program. 25. A process for outputting a keyword group, a process for reconfiguring the keyword group by changing, adding, or deleting the keyword group, and a process for storing the reconfigured keyword group. The automatic sentence classification program according to any one of claims 14 to 24. 26. A computer-readable recording medium in which the automatic document classification program according to claim 14 is recorded. 27. A step of inputting a sentence, a step of analyzing the sentence and extracting a word, and a step of comparing and collating the extracted word with a keyword group in which a keyword is associated with a classification item and described. And a method of weighting the result to find a classification item to which the sentence corresponds. 28. A text analysis step of analyzing a sentence and extracting words, and comparing and collating the extracted words with a keyword group described by associating the extracted words with classification items, and classifying the appearance frequency of the keywords that match the extracted words. An automatic sentence classification method comprising: a keyword scanning step of totaling each item; and a comparison calculation step of weighting the totaled result to find a classification item to which a sentence corresponds. 29. The comparison operation step comprises a step of searching for a classification item corresponding to the matched keyword, a step of multiplying the keyword appearance frequency of each classification item by a coefficient set for each classification item, and weighting. 29. The automatic sentence classification method according to claim 27 or 28, further comprising a step of sorting the classification items by the appearance frequency corrected by the coefficient and totaling the classification results to calculate the classification item to which the sentence corresponds. . 30. A method for automatically classifying a sentence having attribute information, comprising: a text analyzing step of analyzing a sentence having attribute information, extracting attribute information, and extracting a word from a portion other than the attribute information. The extracted words consist of at least two layers of large and small classification items, and the keywords are compared and collated with the keyword group described in association with the small classification items,
The keyword scanning step that aggregates the appearance frequency of the keywords that match the extracted words for each sub-category item, and the large category item to which the sub-category item corresponding to the matching keyword belongs is searched, and the keyword appearance frequency for each sub-category item Of the coefficients set for each large category item for each extracted attribute information, weighting is performed by multiplying by the coefficient of the large category item to which the small category item belongs, and the small category items are sorted by the appearance frequency corrected by the coefficient. And a comparison operation step of calculating a classification item corresponding to a sentence by aggregating the classification results. 31. A method for automatically classifying sentences having attribute information and heading information, analyzing sentences having attribute information and heading information, extracting attribute information, and extracting words from a portion other than the attribute information. A text analysis step, and a keyword scanning step of comparing and collating the extracted word with a keyword group described by associating the keyword with a small classification item, and aggregating the appearance frequency of the keyword that matches the extracted word for each small classification item, The large classification item to which the small classification item corresponding to the matched keyword belongs is searched, and the small classification item belongs to the coefficient set for each large classification item for each attribute information extracted to the keyword appearance frequency for each small classification item. Multiply the weights by the coefficient of the large classification item, sort the small classification items by the appearance frequency corrected by the coefficient, aggregate the classification results, and the sentence corresponds And a comparison calculation step of calculating a classification item to which a sentence corresponds and searching a classification item to which a keyword matching a word extracted from the heading information is calculated. . 32. The keyword scanning step compares and collates the extracted words with at least one or more keyword groups in which the keywords are described by associating the keywords with the code items, and the appearance frequency of the keywords matching the extracted words is determined for each code item. Including the step of totaling, the comparison operation step includes a step of searching for code items corresponding to the matched keywords, sorting the code items by the keyword appearance frequency for each code item, and calculating the code items corresponding to the sentence. 32. The automatic sentence classification method according to claim 27, which is characterized in that. [Claim 33] A keyword group in which a keyword is described in association with a code item includes a prefecture keyword with a prefecture name as a keyword, a country keyword group with a country name as a keyword, and a company keyword group with a company name as a keyword. 33. The automatic sentence classification method according to claim 32, wherein at least one of them is included. 34. The document according to claim 27, wherein the automatic document classification method further comprises a classification result output step of outputting a classification result item calculated by the comparison operation step. Automatic classification method. 35. The automatic document classification method according to claim 27, wherein the automatic document classification method further comprises a conversion result storing step of additionally writing the calculated classification result item into a sentence. . 36. The automatic document classification method further comprises the step of accumulating and / or distributing the classified texts for each item of the classification result. Automatic sentence classification method. 37. The automatic document classification method further collects the classification results for a certain period for each item, statistically processes the appearance frequency of the set keyword, and normalizes the classification result by the number of target sentences for each item. 37. The automatic sentence classification method according to claim 27, further comprising: a classification result statistical step of performing statistical processing of the sentence distribution and a step of outputting the statistical result. 38. A step of outputting a keyword group, a step of changing, adding, or deleting the keyword group to reconstruct the keyword group, and a step of storing the reconstructed keyword group. 38. The automatic sentence classification method according to claim 27. 39. A step of inputting a sentence to an editing terminal,
A step in which the editing terminal sends the sentence to the server system to request a classification process, a step in which the server system receives a sentence and a classification process request, a step in which the server system analyzes the sentence and extracts a word, A step of comparing and matching the extracted word with a keyword group described by associating a keyword with a classification item; a step of weighting the result of the comparison and matching to find a corresponding item;
An automatic sentence classification method, comprising: a step of transmitting a classification processing result to an editor terminal; and a step of the editing terminal receiving the classification processing result.