JPH08314976A - Method and device for retrieving document and document editing device - Google Patents

Abstract

PURPOSE: To reduce the operation quantity which a person inputs a retrieval instructing sentence again to the retrieval instructing sentence which is judged to the unretrievable by a machine. CONSTITUTION: A word dividing part 22 divides an inputted retrieval instructing sentence through the use of a word dictionary. A retrieving processing part 23 obtaines an item index related to the contents of the retrieval instructing sentence and a relation degree expressing the relation degree of the item index by referring to this result, retrieving knowledge and an item index list. A retrieving result judging part 24 by referring to retrieving judging knowledge judges whether the obtained relation degree meets the retrieving judging reference of retrieving judging knowledge or not, and automatically selects retrieval processing to a retrieval instructing sentence which does not meets the reference. When the retrieval instructing sentence meets the reference, a sretrieving result display part 25 displays the item index meeting the retrieving judging reference and attached information belonging to the item index on a display 3. At the time of selecting retrieving processing, a second retrieval instructing sentence preparing part 26 prepares a second retrieval instructing sentence based on the item index and the relation degree by referring to second retrieval instructing sentence preparing knowledge.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to automatic document retrieval,
Document search for selecting an item heading related to the content of the search instruction sentence input by the user through the input device from among a plurality of item headings stored in advance using the search knowledge stored in advance It also relates to document editing to which the document search is applied.

[0002]

2. Description of the Related Art Currently, there is widely researched to classify a text into a specific item from a plurality of items prepared in advance or to search for a specific item by using knowledge created in advance. For example, automatic classification of documents by noun co-occurrence relations automatically extracted from a large amount of document data (IPSJ Research Report Vol. 93, No. 101,
1993). Generally, when searching for an item related to the content of the search instruction text input by the user through the input device, the knowledge that is automatically created from the text representing the content of the item is used as a reference for searching, and The degree of relevance to each item is calculated using the frequency information of the words appearing in and the co-occurrence information between the words, and the item having the high degree of relevance is searched.

[0003]

When the conventional search knowledge is automatically created and the item knowledge related to the content of the search instruction sentence is searched using the search knowledge, the related information item is searched 100% correctly. Is impossible. Therefore, even when the machine determines that the search is impossible, or when the search result is obtained, it is possible for the person to make a determination, input the second search instruction sentence from the search result, and perform the search again. In many cases, there is the problem of putting extra effort on humans. An object of the present invention is to solve the above-mentioned problems, and for a search instruction text that a machine judges to be unsearchable, a document search method and a document search that reduce the amount of work for a human to input the search instruction text again. An object is to provide an apparatus and a document editing apparatus to which the document searching apparatus is applied.

[0004]

In order to achieve the above object, the present invention uses a search knowledge stored in advance to allow a user to select from a plurality of item headings stored in advance via an input device. In the document search method, the item heading related to the content of the search instruction sentence written in natural language consisting of one or more words representing the search instruction desired by the user and input is selected and displayed on the output device. An item heading that is related to the content of the search instruction, and a degree of relevance that indicates the degree of relevance of the item heading, and a process that is stored in advance and that is executed next according to the value of the obtained degree of relevance Refer to the search judgment knowledge that describes the criteria for selecting the content of, and determine whether the search determination criteria of the search determination knowledge are satisfied. If not, re-search processing is automatically performed for the search directive. Choice It has to so that. Further, when the re-search process for the search instruction text is automatically selected, the second search instruction text is created based on the item heading and the degree of association. Further, the second search instruction text is created by referring to the second search instruction text creation knowledge that describes the method of creating the second search instruction text that is stored in advance. Also, word dictionary, search knowledge,
A document search that includes an external storage device that stores a plurality of item headings, a display, an input device, and a processing device, and selects an item heading that is related to the content of the search instruction text input by the user via the input device. In the apparatus, the search determination knowledge and the second search instruction sentence creation knowledge describing the method for producing the second search instruction sentence are stored in the external storage device, and the processing device is input by the user through the input device. The search instruction is word-divided using the word dictionary, and the word division means for outputting the word division result, and the word division result, the search knowledge, and the item heading are referred to and are related to the content of the search instruction. With reference to the item heading, a search processing unit that obtains a degree of relevance that indicates the degree of association of the item heading, and the search determination knowledge, the obtained degree of association is a search determination criterion of the search determination knowledge. Search result determination means that determines whether or not the search is satisfied and automatically selects re-search processing for the search directive that is not satisfied; and if the result of the determination is that the search determination criterion is satisfied, the search determination criterion is satisfied. A second search that describes an item heading, a search result display unit that displays attached information of the item heading on the display, and a method of creating the second search directive when the re-search process is selected. A second search instruction statement creating means for creating a second search instruction statement based on the item heading and the degree of association is provided by referring to the instruction statement creation knowledge. Further, in a document editing device that includes an external storage device storing a document file, a display, an input device, and a processing device, and edits a document file input by a user via the input device, a word dictionary is stored in the external storage device. , A search knowledge, a plurality of item headings, a search determination knowledge, and a second search instruction sentence creation knowledge describing a method for creating a second search instruction sentence are stored, and the user inputs them to the processing device through an input device. The search instruction sentence is divided into words using the word dictionary, and a word dividing unit that outputs the word division result, and the word division result, the search knowledge, and the item heading are referenced to relate the contents of the search instruction sentence. The item heading to be searched, the search processing means for obtaining the degree of association indicating the degree of association of the item heading, and the search determination knowledge are referred to, and the obtained degree of association is the search determination knowledge. Search result determination means for determining whether or not the search determination criteria are satisfied, and automatically selecting re-search processing for search directives that are not satisfied, and a search result if the determination results satisfy the search determination criteria, Item headings that meet the criteria,
Search result display means for displaying attached information of the item heading on the display, and second search instruction sentence creation knowledge describing a method for producing the second search instruction statement when re-search processing is selected. Referring to, a second search instruction statement creating means for creating a second search instruction statement based on the item heading and the degree of association is provided, and related to the content of the search instruction statement input by the user via the input device. The selected item heading is displayed on the display.

[0005]

According to the above means, the search knowledge stored in advance is used to correlate with the content of the search instruction sentence input by the user through the input device from among the plurality of item headings stored in advance. Search results obtained even for search directives in which the degree of association between the content of the search directive and the item header does not satisfy the search criterion when selecting the item header User selects a re-search process automatically using the search instruction sentence input by the user through the input device and the second search instruction sentence created from the search result. It is possible to search for item headings to be performed at once.

[0006]

FIG. 2 shows the hardware configuration of an embodiment of the present invention. The external storage device 1 describes a word dictionary, a word division result file, a search knowledge base file that is search knowledge, a file that describes an item heading list, a file that describes search results, and a search determination knowledge. The file, the file in which the search result after the determination is described, the second search instruction sentence creation knowledge, and the file in which the document database is described are stored. The computer 2 is associated with a search instruction base defined by a user through an input device and described in a natural language consisting of one or more words representing a search instruction desired by the user, from a search knowledge base defined in advance. An item heading is selected, and for the search instruction sentence whose degree of relevance to the selected item heading is determined to be less than the search determination criterion, the related item heading is again determined according to the second search instruction sentence creation knowledge. select. The display 3 displays the result, and the input device 4 inputs the search instruction text by the user.

FIG. 1 is a functional block diagram showing the configuration of an embodiment of the present invention. In FIG. 1, the input device 4
An instruction input unit 21 for inputting a search instruction sentence from a user through the user, a word dictionary 101 for inputting the search instruction sentence into words, and a word dividing unit 22 for outputting the extracted words to the word division result file 102, a word The search processing unit 23 that inputs the division result file 102, the search knowledge base file 103, and the item headline list file 104, calculates the degree of association with each item headline, and outputs the result to the search result file 105, the search result file 105 and the search determination knowledge file 106 are input, it is determined whether or not each item heading of the search result satisfies the search determination criteria according to the search determination conditions, and the result is a search result file 107 after determination.
A search result display unit for inputting the search result judgment unit 24, the post-judgment search result file 107, and the document database file 109 to be output to and displaying the item headings that are the search results and the item contents that are ancillary contents thereof on the display 3. 25, the post-judgment search result file 107 and the second search instruction sentence creation knowledge file 108 are input, a second search instruction sentence is generated according to the contents of the post-judgment search result file 107, and is output to the word division unit. 2 Search instruction sentence creation unit 2
6. The search knowledge creating unit 27 that creates the search knowledge base file 103 and the item heading list file 104 in advance by referring to the document database file 109 composed of the item contents describing the item headings and the contents related to the item headings. , Which are stored as programs in the computer 2, respectively. Also, a word dictionary 101, a word division result file 102, a search knowledge base file 103,
File 104 in which a list of item headings is described, search result file 105, file 106 in which search determination knowledge is described, file 1 in which search results after determination are described
07, the file 108 in which the second search instruction sentence creation knowledge is described, and the document database file 109 are stored in the external storage device 1.

The instruction input unit 21 inputs a search instruction sentence from the user via the input device 4, and outputs the content thereof to the word dividing unit 22. The word division unit 22 inputs the word dictionary 101, divides the search instruction sentence into words, and outputs the extracted words to the word division result file 102. There are already many well-known examples of word segmentation processing (for example, the 44th National Conference of Information Processing Society of Japan, 3-181, '92 .3), so the explanation is omitted here. Search processing unit 23
Inputs the word division result file 102, the search knowledge base file 103, and the item headline list file 104, calculates the degree of association with each item headline, and outputs the result to the search result file 105. The search result determination unit 24 inputs the search result file 105 and the search determination knowledge file 106, determines whether or not each item heading of the search result satisfies the search determination criterion according to the search determination condition, and determines the result. It is output to the post-search result file 107. The search result display unit 25 displays the search result file 1 after the determination.
07 and the document database file 109 are input, and the item headings that are the search results and the item contents that are the ancillary contents thereof are displayed on the display 3. The second search instruction sentence creating unit 26 determines the post-judgment search result file 107, and
The second search instruction sentence creation knowledge file 108 is input, a second search instruction sentence is generated according to the contents of the post-judgment search result file 107, and the second search instruction sentence is output to the word division unit. The search knowledge creation unit 27 refers to the document database file 109 composed of a plurality of item headings and the item contents describing the contents related to each item heading, and searches the search knowledge base file 103 and the item headline list file 104. To create.

FIG. 3 shows a screen example 1 on the display 3. The screen on the display 3 has, on the displayable area 31, an area 32 for inputting the current search target statement, an area 33 for displaying the post-judgment search result file 107, and an item heading that is an answer to the search target text. Is displayed as an answer sentence. This screen is a screen when the answer has not been decided yet, and the answer is not displayed in the answer sentence column.

FIG. 4 shows an overall processing flow chart in this embodiment. The processing of this embodiment will be described with reference to FIG. First, the instruction input unit 21 inputs a search instruction sentence from the user via the input device 4, and outputs the content thereof to the word dividing unit 22 (step 201). Next, the word dividing unit 22 inputs the word dictionary 101, divides the search instruction sentence into words, and extracts the extracted words from the word division result file 10.
2 (step 202). Next, the search processing unit 2
3, the word division result file 102, the search knowledge base file 103, and the item heading list file 1
Enter 04 to calculate the degree of association for each item heading,
The result is output to the search result file 105 (step 203). Next, in the search result determination unit 24, the search result file 105 and the search determination knowledge file 106 are input, it is determined whether or not the search result satisfies the search determination criterion, and only those satisfying the criterion are determined and the search result is determined. Output to the file 107 (step 204). Next, when the search processing ends in step 204, the search result display unit 25 displays the post-judgment search result file 107, and
The document database file 109 is input and the search result is displayed on the display 3 as the answer sentence of the search instruction sentence (step 205). Next, when the search processing is not completed in step 204, the second search instruction statement creating unit 26 inputs the post-judgment search result file 107 and the second search instruction statement creation knowledge file 108, and after the decision According to the contents of the search result file 107, a second search instruction sentence that combines the search instruction sentence input by the user and the search result is generated, and the process returns to step 202 (step 206).

FIG. 5 shows the file structure of the word dictionary 101. The word dictionary 101 is composed of a plurality of records 1011 and each record is a “headword”,
It is one storage unit of data composed of two data items 1012 of “part of speech”. For example, the word dictionary 10 in FIG.
In 1, the word "part of speech" with the word "headword" of "Heian period" is "noun", and the word "part of speech" with the word "headword" of "serving" is "verb".
It indicates that the word “part of speech” with the word “entry word” of “essay” is a “noun”. This word dictionary 101
Is prepared in advance.

FIG. 6 shows the file structure of the word division result file 102. The word division result file 102 is composed of a plurality of records 1021, and each record is one storage unit of data composed of three data items 1022 of “word”, “part of speech”, and “appearance frequency”. For example, in the word division result file 102 shown in FIG. 6, the search instruction sentence input by the user indicates “Heian period”,
"Miyanaka", "serving", "essay", "work", "author"
Is extracted, and the "part of speech" of the "word" called "Heian period" is "noun", which means that the word appears once in the search instruction sentence. The word division result file 102 is a file created by the word division unit 22.

FIG. 7 shows the file structure of the search knowledge base file 103. The search knowledge base file 103 is composed of a plurality of records 1031.
Each record is a storage unit of data composed of four data items 1032 of “item heading”, “keyword”, “part of speech”, and “weight”. For example, in the search knowledge base file 103 of FIG. 7, it is shown that the “item heading” is “Makurasoshi”, the “keyword” is “Heian period”, the “part of speech” is “noun”, and the “weight” is “2”. ing. The search knowledge base file 103 is created in advance by the search knowledge creating unit 27.

FIG. 8 shows the file structure of the item heading list file 104. The item heading list file 104 is composed of a plurality of records 1041.
Each record is one storage unit of data composed of the data item 1042 of “item heading”. For example, in FIG.
In the item heading list file 104 of
It indicates that there are "item headings" such as "The Tale of Genji", "The Story of Heike", and "Seishonagon". This item headline list file 104 is stored in advance in the search knowledge creation unit 27.
It was created in.

FIG. 9 shows the file structure of the search result file 105. Search result file 105
Is composed of a plurality of records 1051, and each record is one storage unit of data composed of data items 1052 of “search instruction sentence”, “item heading”, and “degree of association”. For example, in the search result file 105 shown in FIG.
When the "search directive" is "the author of the work that wrote the experience of serving in the court during the Heian period", the "relevance" is "130" for the item whose "item heading" is "Makurasoshi"
"Item heading" is related with the value "Tomagariga", "relevance" is related with the value "120", "item heading" is "purple Shikibu diary"
"Relevance" is associated with the value "105" for items that are "," Relevance "is associated with the value" 80 "for items whose" Item heading "is" Tosa Diary " It indicates that This search result file 105
Is a file created by the search processing unit 23.

FIG. 10 shows the file structure of the search judgment knowledge file 106. The search determination knowledge file 106 is composed of a plurality of records 1061, and each record is composed of three data items 1062 of “relevance condition”, “search result file description content”, and “processing content”. It is a memory unit. For example, FIG.
In the search determination knowledge file 106 of 0, when the item “relevance degree condition” is “one item heading with a degree of relevance value of 100 or more exists”, the “description content” is “search instruction sentence, relevance degree”. It indicates that "the item heading and the degree of relevance satisfying the condition (1)" and "the processing content" are "search result display and search processing end". When the item “relevance degree condition” is “two or more item headings with relevance value of 100 or more exist”, “description content” is “search instruction sentence, item heading condition relevance condition”. , Relevance ”, and“ processing content ”is“ re-search processing ”. Further, when the item of "relevance degree condition" is "more than 0 item headings with relevance value of 100 or less exist", "description content" is "search instruction sentence, no item heading", and " “Processing content” indicates “prompt the user to re-enter detailed search instruction text”. The content of the search determination knowledge file 106 can be changed by the user, and the user can change the content by inputting the description content of the search determination knowledge file 106 via the input device.

FIG. 11 shows the post-judgment search result file 107.
Shows the file structure of. The post-judgment search result file 107 is composed of a plurality of records 1071, and each record includes a “search instruction sentence”, an “item heading”,
It is a storage unit of data composed of three data items 1072 of “degree of association”. For example, in the post-judgment search result file 107 of FIG. 11, when the “search directive” is “the author of the work that wrote the experience of serving in the court during the Heian period”, the “item heading” is “Makurasoshi” "Relevance" is related to a certain item with a value of "130",
An item whose "item heading" is "Togagarakari" is associated with a value of "120" for "relevance",
For items whose "Item Heading" is "Murasaki Shikibu Diary",
The value "105" indicates that the "relevance" is related. The difference from the search result file 105 is that there is no description of item headings that do not satisfy the search determination criteria. The post-judgment search result file 107 is a file created by the search processing unit 23.

FIG. 12 shows the file structure of the second search instruction sentence creation knowledge file 108. The second search instruction sentence creation knowledge file 108 includes a plurality of records 10
Each record is a storage unit of data composed of two data items 1082 of “condition of re-search method” and “search method”. For example, in the second search instruction sentence creation knowledge file 108 of FIG. 12, the item “condition of re-search method” is “the deviation value of the highest degree of association is 70”.
When it is "above", the "search method" indicates that "the item heading having the highest value of the degree of association is the second search instruction sentence". Further, when the item of “condition of re-search method” is “the deviation value of the highest degree of association is less than 70”, the “search method” is set to the item heading having the value of the highest degree of association and the search knowledge file. The keyword of the described applicable item heading is used as the second search instruction sentence. " The contents of the second search instruction sentence creation knowledge file 108 can be changed by the user, and the user inputs the description contents of the second search instruction sentence creation knowledge file 108 via the input device to change the contents. Can be changed.

FIG. 13 shows the document database file 10.
9 shows the file structure of No. 9. The document database file 109 is composed of a plurality of records 1091, and each record is one storage unit of data composed of two data items 1092 of “item heading” and “item content”. For example, in the document database file 109 of FIG. 13, when the item in the “item heading” is “Makurasoshi”, the “item content” is “an essay in the Heian period. It is a summary. " The document database file 109 is prepared in advance.

FIG. 14 is a processing flow of the instruction input unit 21. The processing of the instruction input unit 21 will be described with reference to FIG. First, when the user inputs a search instruction sentence via the input device 4, the input content is output to the word dividing unit (step 2101). Next, the search processing unit 23, and
The series of searches in the second search instruction sentence creating unit 26 is completed, and the input waiting state is entered until a signal permitting the input of the next search instruction sentence is input from the search processing unit 23 (step 210).
2). Next, when the input permission signal is input, the input waiting state is entered until the user inputs the search instruction text via the input device 4 (step 2103).

FIG. 15 is a processing flow of the word dividing unit 22. The process of the word dividing unit 22 will be described with reference to FIG. First, the search instruction sentence and the word dictionary 101 are input (step 2201). Next, the search instruction sentence is divided into words, and the extracted words are output to the word division result file 102 (step 2202). Next, the fact that the word division is completed is output to the search processing unit 23 (step 220).
3).

FIG. 16 is a processing flow of the search processing section 23. The processing of the search processing unit 23 will be described with reference to FIG. First, the word division result file 102, the search knowledge base file 103, the item heading list file 10
4 is input (step 2301). Next, the degree of association with the search directive is calculated for every item heading described in the item heading list file. As for the method of calculating the degree of association of each item heading, if each “keyword” of the corresponding “item heading” of the search knowledge base file 103 matches the “keyword” of the word division result file 102,
The sum of the products of the "weight" and "appearance frequency" of the "keyword" is calculated. Then, the deviation value of the total value calculated for each item heading is set as the degree of association of the item heading (step 2302). Next, the calculated relevance is output to the search result file 105 in descending order of relevance (step 23).
03). Next, the fact that the processing in the search processing unit 23 is completed is output to the search result determination unit 24 (step 230).
4).

FIG. 17 is a processing flow of the search result judging section 24. The processing of the search result determination unit 24 will be described with reference to FIG. First, the search result file 105 and the search determination knowledge file 106 are input (step 240).
1). Next, determination of search is performed. The search determination method is as follows. When the content described in the search result file 105 satisfies the “relationship condition” of the search determination knowledge file 106, the post-judgment search result file 107 is determined according to the “search result file description content”. It is described and the following processing is performed according to the "processing content" (step 2402).

FIG. 18 is a processing flow of the search result display unit 25. The processing of the search result display unit 25 will be described with reference to FIG. First, the post-judgment search result file 107
Is input (step 2501). Next, the search instruction sentence currently targeted is displayed in the area 32 of the display 3 of FIG. 3 (step 2502). Next, the area 33
After that, the search result file 107 is displayed (step 2503). Next, in the area 34, when one item heading to be the answer to the search instruction sentence is determined, this item heading is displayed as the answer sentence (step 2504).

FIG. 19 is a processing flow of the second search instruction sentence creating section 26. The process of the second search instruction sentence creating unit 26 will be described with reference to FIG. First, the post-judgment search result file 107 and the second search instruction sentence creation knowledge file 108
Is input (step 2601). Next, when the content described in the post-judgment search result file 107 satisfies the “condition of the re-search method” of the second search instruction sentence creation knowledge, the second content is searched according to the content of the “search method”. A search instruction sentence is generated (step 2602). Next, the second generated
Is output to the word division unit 21 (step 2
603).

FIG. 20 is a processing flow of the search knowledge creating unit 27. The process of the search knowledge creating unit 27 will be described with reference to FIG. First, the document database file 10
9 is input (step 2701). Next, item headings are extracted from the document database file 109 and output to the item heading list file 104 (step 270).
2). Next, the keywords, parts of speech, and appearance frequencies extracted for each item heading from the document database file 109 are output to the search knowledge base file 103 (step 276).
03).

Next, the above-mentioned series of processing will be described for the case where the search instruction sentence shown in FIG. 3 "Who is the author of the work that wrote the experience of serving in the court during the Heian period?" FIG. 6 uses the word dictionary 101 shown in FIG.
This is the word division result file 102 output as a result of word division. That is, it indicates that words such as “Heian period”, “serving”, and “essay” have been extracted from this search instruction sentence. Further, FIG. 9 shows the word division result file 102 as the search knowledge base file 103 shown in FIG.
It is a search result file 105 searched by using the item heading list file 104 shown in FIG. In addition, FIG.
Is the post-judgment search result file 107 showing the result of judgment using the search judgment knowledge file 106 shown in FIG.
Is. In the search result file 105 of FIG. 9, there are three item headings having a degree of association of 100 or more. Therefore,
In the post-judgment search result file 107, the search instruction text, the three item headings and the degree of association are described, and the display instruction text and the three item headings are displayed as the post-judgment search results on the display 3. No answer is displayed in the answer text column because only one answer has been determined. Further, as the next processing content, re-search processing is performed.

Next, in the second search instruction sentence creating section 26, as the condition of the second search instruction sentence producing knowledge, the highest value among the deviation values of these three degrees of association is 61, so the search method is , “Set item heading and keyword with highest relevance value as second search directive”. Therefore, as the second search instruction text, "Makura-soshi", "Heian period", "serving", "essay", etc. are the second search instruction texts. Then, the second search instruction sentence is output to the word dividing unit 22. The result of further search by the search processing unit is shown in the search result file after the second search in FIG. In the search result file of FIG. 21, the item heading with the degree of association of 100 or more is only one of “Seishonagon”. Therefore, the second search result file in FIG.
The search instruction sentence, the item heading, and the degree of association are described, and the result is displayed on the display. Since only one answer has been determined this time, the answer is displayed in the field of the answer sentence, and the series of search processing ends. FIG. 23 is a screen example of a result of the recursive search.

As described above, according to the present embodiment, the search instruction input by the user through the input device 4 is selected from the plurality of item headings stored in advance using the search knowledge stored in advance. When selecting an item heading that is related to the content of a sentence, even if the degree of relevance that indicates the degree of association between the content of the search heading and the item heading does not satisfy the search determination criteria, , The re-search process is automatically selected using the obtained search result, and the search is performed again using the search instruction sentence input by the user and the second search instruction sentence created from the search result. It is possible to search for the desired item heading at once.

Next, an extended example of the present invention will be described. FIG. 24 is an overall configuration diagram of a document editing apparatus with a dictionary reverse search function, which is an example of this extension. In FIG. 24, the input device 4
Document input unit 2 for inputting a document input by the user via the
8, a document file 111 stored in advance, a display unit 29 that edits a document input by the user via the input device 4 and displays it on the display 3, a word dictionary 101, a search for an input sentence input by the user Knowledge base 103,
Item headline list 104, search determination knowledge 106, second search instruction sentence creation knowledge 108, dictionary database file 11
The reverse dictionary search unit 2a that outputs an item heading related to the content of the input sentence with reference to 0, and the temporary storage unit 2b that temporarily stores the document file edited on the display unit 29 are stored in the computer 2 as a program. Has been done. The dictionary reverse search unit 2a performs the series of recursive search processing described in this embodiment, and outputs the result to the document editing unit 29. Further, this allows the user to acquire the item heading corresponding to the meaning content described in the text. Thus, even when the user wants to check the meaning, reading, and example while editing a document, the search result can be used as it is for the document being edited. This extended example
This is effective when applied to a document editing device.

[0031]

According to the present invention, the contents of the search instruction sentence input by the user via the input device 4 from the plurality of prestored item headings using the prestored search knowledge. When selecting an item heading that is related to, the degree of relevance that indicates the degree of relevance between the content of the search heading and the item heading is obtained even for a search heading that does not satisfy the search determination criteria. The user selects a re-search process automatically by using the retrieved search result, and the second search directive created by the user and the second search directive created from the search result are used to perform the search again. It is possible to search for item headings to be performed at once.

[Brief description of drawings]

FIG. 1 is a functional block diagram showing a configuration of an exemplary embodiment of the present invention.

FIG. 2 is a diagram showing a hardware configuration of an embodiment of the present invention.

FIG. 3 is a diagram showing an example of a screen on a display.

FIG. 4 is a flowchart of the overall processing in the embodiment.

FIG. 5 is a file configuration diagram of a word dictionary.

FIG. 6 is a file configuration diagram of a word division result file.

FIG. 7 is a file configuration diagram of a search knowledge base file.

FIG. 8 is a file configuration diagram of an item heading list file.

FIG. 9 is a file configuration diagram of a search result file.

FIG. 10 is a file configuration diagram of a search determination knowledge file.

FIG. 11 is a file configuration diagram of a post-judgment classification result file.

FIG. 12 is a file configuration diagram of a second search instruction sentence creation knowledge file.

FIG. 13 is a file configuration diagram of a document database file.

FIG. 14 is a diagram showing a processing flow of an instruction input unit.

FIG. 15 is a diagram showing a processing flow of a word dividing unit.

FIG. 16 is a diagram illustrating a processing flow of a search processing unit.

FIG. 17 is a diagram illustrating a processing flow of a search result determination unit.

FIG. 18 is a diagram showing a processing flow of a search result display unit.

FIG. 19 is a diagram showing a processing flow of a second search instruction sentence creating unit.

FIG. 20 is a diagram showing a processing flow of a search knowledge creating unit.

FIG. 21 is a diagram showing a file example of a search result file after the second search.

FIG. 22 is a diagram showing a file example of a post-judgment search result file after the second search.

FIG. 23 is a diagram showing another example of a screen on the display.

FIG. 24 is a functional block diagram of a document editing device with a dictionary reverse search function that is an extension example of the present invention.

[Explanation of symbols]

 1 External Storage Device 2 Computer 2a Reverse Dictionary Search Unit 2b Temporary Storage Unit 3 Display 4 Input Device 101 Word Dictionary 102 Word Division Result File 103 Search Knowledge Base File 104 Item Headline List File 105 Search Result File 106 Search Judgment Knowledge File 107 After Judgment Search result file 108 Second search instruction sentence creation knowledge file 109 Document database file 110 Dictionary database 111 Document file 21 Instruction input section 22 Word division section 23 Search processing section 24 Search result determination section 25 Search result display section 26 Second search instruction statement Creation unit 27 Search knowledge creation unit 28 Input unit 29 Display unit

Claims (9)

[Claims] 1. One or more words representing a search instruction desired by the user, which is input by the user through an input device, from a plurality of item headings stored in advance using search knowledge stored in advance. In a document search method for selecting and displaying an item heading related to the content of a search instruction written in natural language on an output device, an item heading related to the content of the search instruction and an item The degree of relevance indicating the degree of relevance of headings is obtained, and the search determination knowledge describing the criteria for selecting the content of the process to be executed next, which is stored in advance, is referred to according to the value of the obtained degree of relevance. A document search method characterized by determining whether or not a search determination criterion of the search determination knowledge is satisfied, and if not, automatically selecting a re-search process for a search instruction sentence. 2. The document search method according to claim 1, wherein when the user inputs a change in the description content of the search determination knowledge via an input device, the description content of the search determination knowledge stored in advance is stored. A document retrieval method characterized by changing. 3. The document search method according to claim 1, wherein when a re-search process for the search instruction sentence is automatically selected, a second search instruction sentence is created based on the item heading and the degree of association. A document retrieval method characterized by. 4. The document search method according to claim 3, wherein the second search instruction statement refers to second search instruction statement creation knowledge that describes a method of creating a second search instruction statement stored in advance. A document retrieval method characterized by being created by. 5. The document search method according to claim 4, wherein when the user inputs a change in the description content of the second search instruction sentence creation knowledge through an input device, the second search stored in advance. A document retrieval method characterized by changing the description content of instruction sentence creation knowledge. 6. A content of a search instruction sentence input by a user via an input device, comprising a word dictionary, search knowledge, an external storage device storing a plurality of item headings, a display, an input device, and a processing device. In a document search device for selecting related item headings, the external storage device stores search determination knowledge and second search instruction sentence creation knowledge describing a method of creating a second search instruction sentence, and the processing device Refers to the word division means for dividing the search instruction sentence input by the user through the input device into words using the word dictionary and outputting the word division result, the word division result, the search knowledge, and the item heading. hand,
An item heading related to the content of the search instruction sentence, a search processing unit that obtains a degree of relevance that indicates the degree of relevance of the item heading, and the retrieval degree is referred to by the retrieval processing unit. Search result determination means for determining whether or not the search determination criterion of the determination knowledge is satisfied, and automatically selecting re-retrieval processing for the search instruction statement that is not satisfied; and the result of the determination, when the search determination criterion is satisfied Is an item heading that satisfies the search judgment criteria, a search result display means for displaying the attached information of the item heading on the display, and creates the second search instruction sentence when re-search processing is selected. A document inspection comprising a second search instruction statement creating means for creating a second search instruction statement based on the item heading and the degree of association with reference to the second search instruction statement creation knowledge describing the method. Apparatus. 7. The document search device according to claim 6, further comprising a document database including item headings and item contents, wherein the processing device reads out the item headings and item contents of the document database to obtain the search knowledge. A document search device comprising a search knowledge creating means for creating. 8. The document search device according to claim 6, wherein the processing device changes the description content of the search determination knowledge or the second search instruction sentence creation knowledge into a change input input by a user via an input device. Therefore, the document retrieval apparatus is provided with a changing unit. 9. A document editing apparatus, comprising an external storage device storing a document file, a display, an input device, and a processing device, for editing a document file input by a user via the input device. A word dictionary, a search knowledge, a plurality of item headings, a search determination knowledge, and a second search instruction sentence creation knowledge in which a method for creating a second search instruction sentence is described are stored in the processing device. Word division of the search instruction sentence input via the word dictionary using the word dictionary, the word division means for outputting the word division result, with reference to the word division result, search knowledge and item heading,
An item heading related to the content of the search instruction sentence, a search processing unit that obtains a degree of relevance that indicates the degree of relevance of the item heading, and the retrieval degree is referred to by the retrieval processing unit. Search result determination means for determining whether or not the search determination criterion of the determination knowledge is satisfied, and automatically selecting re-retrieval processing for the search instruction statement that is not satisfied; and the result of the determination, when the search determination criterion is satisfied Is an item heading that satisfies the search judgment criteria, a search result display means for displaying the attached information of the item heading on the display, and creates the second search instruction sentence when re-search processing is selected. A second search instruction statement creating means for creating a second search instruction statement based on the item heading and the degree of association is provided by referring to the knowledge of creating the second search instruction statement that describes the method, and the user uses the input device. A document editing apparatus, which selects an item heading related to the content of a search instruction sentence input by a user and displays it on a display.