JP4046221B2 - Document processing device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
  The present inventionDocument processing deviceIn detail, it helps the user to find out the conceptual expression that the user needs by searching and narrowing down the conceptual expression contained in the document or document collectionDocument processing device.
[0002]
[Prior art]
  Recently, computerization of information has progressed, and documents conventionally stored in paper documents have also been digitized. Along with the digitization of such documents, a large number of digitized documents are distributed, and how to manage and easily reuse collected and accumulated digitized documents has become an important issue. In this case, an analysis technique for finding some knowledge from a large amount of electronic documents is desired. This analysis technology is a technology that grasps the contents of each group of documents while grouping a large number of documents by document content. Conventionally, user-driven input of group definition formulas and search technology There have been proposed a method of grouping by using a method of automatically grouping using a clustering technique based on the co-occurrence of words in a document.
[0003]
  Conventionally, as a grouping technique, for example, “Data analysis system” of Patent Document 1 has been proposed, and this data analysis system extracts a concept (keyword in a phrase) from a text and specializes it for the purpose. A category dictionary (thesaurus) is used to convert the expression in the document into labeled data. In addition, the dependency relationship between concepts is analyzed, and a characteristic concept is extracted from the concept frequency / cross table as a concept combination.
[0004]
  Conventionally, a document processing apparatus disclosed in Patent Document 2 has been proposed. This document processing apparatus stores a morpheme analysis unit that performs morpheme analysis on an input sentence, a specific expression candidate acquisition unit that uses a partial sequence of a morpheme sequence as a specific expression candidate with a weight, and stores some specific expressions in advance The specific expression dictionary and a specific expression dictionary search unit for outputting a real number representing a matching degree of the expression in the specific expression dictionary of the morpheme sequence as a search result of the specific expression dictionary, and the specific expression candidate given to the candidate A discriminant score is calculated by using the weight and the search result of the candidate for the specific expression dictionary as a variable, and a discriminant analysis execution unit that excludes candidates whose discriminant score is less than a certain value; An output unit that outputs a morpheme character string that is not excluded by the discriminant analysis execution unit as a specific expression, calculates a discriminant score, and determines whether to leave it as a specific expression candidate , It is intended to make an accurate judgment.
[0005]
  That is, this prior art performs word analysis and dependency analysis from text, groups sentences according to the similarity of sentence structures, and correlates from the correlation between the number of appearances of keywords extracted from text and grouped sentences. Extracting strong items.
[0006]
[Patent Document 1]
JP 2001-75966 A
[Patent Document 2]
JP 2000-172691 A
[0007]
[Problems to be solved by the invention]
  However, in such a conventional technique, in order to allow a user to freely search for and discover a concept included in a document or a document collection, the discovery of a concept important for analyzing document data can be performed. There was a need for improvement.
[0008]
  That is, in the prior art described in Patent Document 1, a concept is extracted from text, a category label is attached based on a category dictionary, a dependency relationship between concepts is analyzed, and a combination of concepts is also a concept. Furthermore, since characteristic concepts are extracted by concept frequency / cross table, it is necessary to create a category dictionary in advance, and there is a problem that it takes a burden to create and maintain the category dictionary, and frequency information Therefore, there is a problem in that it cannot be extracted unless the number of appearances is large.
[0009]
  Moreover, in the prior art described in Patent Document 2, word analysis and dependency analysis are performed from text, and sentences are grouped according to the similarity of sentence structures, and keywords extracted from the text and grouped sentences are compared. Since items having a strong correlation are extracted from the correlation of the number of appearances, there is a problem that the items cannot be extracted unless the number of appearances is large.
[0010]
  However, in the analysis of document data, especially in the analysis of data such as questionnaires, even a concept that is not characteristic, for example, a concept that is not frequent, may be a concept that a user needs for analysis. In the case of such a concept, whether or not it is necessary depends on the user's intention and purpose, and it is difficult to automatically extract it. Therefore, in order to support the discovery of such a concept, it is necessary to assist the user in freely searching for and discovering the concept included in the document or document set.
[0011]
  Therefore, the present invention supports a user to search a conceptual expression included in a document or a document set while performing search or narrowing down and find a conceptual expression required by the user.Provide a document processing deviceThe purpose is that.
[0012]
[Means for Solving the Problems]
  The invention according to claim 1 is a language analysis unit that performs morphological analysis and dependency analysis on input document data, and a document that holds language information of the document data based on a language analysis result in the language analysis unit. Generate data structures and in clausesWhen a specific independent word or ancillary word appears inFor words and phrases in the document data structure, Indicate cancellations, requests, questions or possibilities depending on the specific independent words or appendices that appearDocument data structure generation means for assigning a semantic tag, document data structure storage means for storing the document data structure with the semantic tag generated by the document data structure generation means, and an arbitrary number designated by the user Of words and phrases in the document data structureMeaning tagA concept expression specifying means for receiving a concept expression consisting of a combination of the above and a concept expression related to the concept expression received by the concept expression specifying means from the document data structure stored in the document data structure storage means The document processing apparatus is characterized by comprising a concept expression extracting means.
[0013]
  The invention of claim 2 further comprises a concept expression display means for displaying a concept expression to be selected and specified based on the document data structure, the concept expression specifying means from the concept display displayed on the concept expression display means. 2. The document processing apparatus according to claim 1, wherein designation of an arbitrary concept expression is received, and the concept expression display means displays a concept expression extraction result by the concept expression extraction means.
[0014]
  The invention of claim 3 further comprises a synonym dictionary, and the document data structure generation means generates the document data structure by adding representative notation information to synonym / notation words based on the synonym dictionary. The document processing apparatus according to claim 1, wherein the document processing apparatus is a document processing apparatus.
[0015]
  The invention of claim 4 further includes means for holding a history of concept expressions received by the concept expression designating means, and based on the history, a concept expression extraction result based on a concept expression designation made in the past is obtained as the concept expression. The document processing apparatus according to claim 2, wherein the document processing apparatus re-displays the display means.
[0016]
  The invention according to claim 5 is characterized in that the concept expression display means displays the document data including the concept expression received by the concept expression specifying means or a part of the document data. The document processing apparatus according to any one of the above items.
[0017]
  According to the invention of claim 6, the concept expression extracting means conforms to a concept expression (hereinafter referred to as designated concept expression) in which the document data structure stored in the document data structure storage means is received by the concept expression specifying means. And extracting from the matching document data structure a word in the specified conceptual expression and a dependency-related word, extracting a semantic tag not in the specified conceptual expression, and extracting the extracted word and the semantic tag The document processing apparatus according to claim 1, wherein a conceptual expression including the above is created.
[0018]
  DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, preferred embodiments of the invention will be described in detail with reference to the accompanying drawings. The embodiments described below are preferred embodiments of the present invention, and thus various technically preferable limitations are given. However, the scope of the present invention is particularly limited in the following description. As long as there is no description which limits, it is not restricted to these aspects.
[0019]
  1 to 14 are diagrams showing an embodiment of a document processing apparatus, a document processing method, and a recording medium according to the present invention. FIG. 1 is an example of a document processing apparatus, a document processing method, and a recording medium according to the present invention. It is a block block diagram of the document processing apparatus 1 to which embodiment is applied.
[0020]
  In FIG. 1, the document processing apparatus 1 includes a document input unit 2, a language analysis unit 3, a document data structure generation unit 4, a document data structure storage unit 5, a concept expression extraction unit 6, a concept expression designation unit 7, and a concept expression display 8. And a recording medium such as a CD-ROM (Compact Disc Read Only Memory) for recording a document processing program and necessary data is read by a computer or the like and introduced.
[0021]
  The document input unit (document input means) 2 inputs a document to be processed, and can input a plurality of documents. When inputting a plurality of documents, an identifier is assigned to each document. And stored in a storage unit or the like. In the following description, it is assumed that a document set is input from the document input unit 2.
[0022]
  The language analysis unit (language analysis means) 3 is a morpheme analysis process for analyzing the morpheme of each document set input from the document input unit 2, a dependency analysis process for analyzing the dependency relationship between phrases of the document to be analyzed, etc. The linguistic attributes obtained by the morphological analysis process, the dependency analysis process, and the like are output to the document data structure generation unit 4 as an analysis unit. Specifically, the language analysis unit 3 analyzes words included in each document of the document set in the morphological analysis process, and analyzes sentences and clauses included in the document in the dependency analysis process. Analyzes clauses that are in the relationship between dependency and reception. For example, when analyzing the sentence “Software installation cannot be executed normally”, the language analysis unit 3 performs dependency analysis after performing morphological analysis as shown in FIG. FIG. 2 shows an example of the analysis result of the above example, where word separators are represented by “/”, “self” above each word represents an independent word, and “attached” represents an attached word. ing. That is, in FIG. 2, the independent word “software” is related to the phrase 1 with the attached word “no”, and the independent word “install” is given the phrase 2 with the attached word “ga”. The phrase 3 with the normal word “normal” and the word “ni” attached is related, and the word “execution” has two additional words “done” and “no”. It shows that the phrase 4 is received and the phrase 2 is also related to the phrase 4 as a dependency.
[0023]
  The document data structure generation unit (document data structure generation means) 4 converts each document in the document set into a data structure as shown in FIG. 3 based on the analysis result of the language analysis unit 3. Information as shown in FIG. For example, the document data structure generation unit 4 generates a word list with unique identifiers for words included in a document or document set as shown in FIG. 5 and manages the words. The part-of-speech information, the appearance frequency in the whole or the number of appearance documents are calculated and added.
[0024]
  That is, the converted data structure shown in FIG. 3 for each document in the document set is as shown in FIG. 4. The document manages a sentence ID list included in the document, and the sentence is its own sentence ID. To manage the phrase list included in the sentence. The phrase manages its own phrase ID, a word ID list included in the phrase, a related phrase ID list, and a received phrase ID. This word ID is an ID in the word list shown in FIG. 5, and the related phrase ID list is an ID of a related phrase that receives the relevant phrase. Since a plurality of clauses can be related clauses for one received clause, they are managed by the related clause ID list. The received clause ID is an ID of a received clause in which the relevant clause becomes a related clause, and the related clause can take only one received clause.
[0025]
  In addition, the document data structure generation unit 4 can also hold the type of dependency relationship as information managed by the clause, for example, whether it is a combination modification or a combination modification, etc. You can also describe the type of relationship by type.
[0026]
  Further, the document data structure generation unit 4 has a synonym dictionary and can have representative notation information regarding words having synonyms, and has synonym representative notations as items in the word list as shown in FIG. Can be realized.
[0027]
  Further, the document data structure generation unit 4 assigns a semantic tag representing an additional meaning to a word or phrase in the document data structure from an attached word expression or the like in the phrase, and a concept expression specifying unit 7 or a concept expression extraction. Part 6,Conceptual expression display section 8In addition, not only words but also semantic tags can be used as concept expressions. This meaning tag is used to add a meaning to a clause as a tag when an attached word or the like in the clause represents a specific additional meaning. For example, the meaning tags “cancellation”, “request”, “possible”, and “question” are added to the phrase when the following words appear in the phrase, and a plurality of phrases are added to one phrase. Semantic tags may be attached.In the following description, an intention tag is synonymous with a semantic tag.
[0028]
  Intent tag ID 1 “cancellation”: auxiliary verb “none”, auxiliary verb “z”, auxiliary verb “mai”, auxiliary auxiliary verb “difficult”, adjective “none”
  Intent tag ID 2 “Request”: auxiliary verb “I want”, verb “I want”, connected particle “te” + verb “I want”
  Intent tag ID 3 “question”: final particle “ka”, final particle “ka” + final particle “na”, symbol “?”
  Intent tag ID 4 “possible”: auxiliary verb “can”, auxiliary verb “re”, auxiliary verb “re”
In the conceptual expression, for example, the meaning tag is represented by an expression “(+ cancellation + possible)”. The meaning tag can be used alone or as a conceptual expression, or can be used in a form added to a word such as “execution (+ possible + cancellation)”.
[0029]
  The document data structure storage unit (document data structure storage unit) 5 stores and manages the document data structure generated by the document data structure generation unit 4.
[0030]
  The concept expression specifying unit (concept expression specifying means) 7 includes an input unit for characters / symbols, a pointing device such as a mouse, and the like, and the user specifies the concept expression. The concept expression display unit (concept expression display means) 8 is a CRT (Cathode Ray Tube), an LCD (Liquid Crystal Display), or the like. The concept expression extraction unit 6 extracts the concept expression and the concept expression designation unit 7. Various data necessary for the document processing device 1 to process the document, such as a concept expression designation screen (concept browser), is displayed.
[0031]
  As a method of designating the concept expression in the concept expression designating unit 7, for example, a direct entry method in which an input dialog as shown in FIG. 6 is displayed on the concept expression display unit 8 and the user directly enters the concept expression, FIG. As shown in FIG. 12, a concept display screen (concept browser) is displayed on the concept expression display unit 8, and a selection method for selecting a concept expression designated by a pointing device such as a mouse from the displayed concept expressions. Can be used. In the example of FIGS. 7 and 8, the word list information is displayed as the initial state, and the word that the user wants to specify (in FIG. 7, “Receive” is selected) is selected. 7 and 8, when a word of concept expression is selected and the “narrow down” button is operated, concept expression extraction is executed on the right side. In this way, when using the selection method, the concept expression display unit 8 can repeatedly search the concept expression while selecting the concept display to be designated next, rather than the direct entry method in the case of FIG. It is possible to specify a conceptual expression efficiently.
[0032]
  The concept expression extraction unit (concept expression extraction means) 6 extracts a concept expression (word (independent word) or intention tag) having a strong relationship with the concept expression specified by the concept expression specifying unit 7 from each document in the document set. The frequency is calculated and the concept expression extraction result is sent to the concept expression display unit 8 for display.
[0033]
  Next, the operation of the present embodiment will be described. The document processing apparatus 1 according to the present embodiment performs document processing based on a concept display according to user designation.
[0034]
  First,This embodimentThe concept expression, which is the basic concept of, will be explained. Documents handled in the present embodiment are basically expressed in Japanese sentences, and conceptual expressions are expressed in units of words (independent words). A single word represents a concept, and a plurality of words represent a concept. For example, the following conceptual expression is used.
[0035]
  1) Search
  2) Information ⇒ Search
  3) Information ⇒ Search ⇒ Service
  4) Software ⇒ Install (+ possible + cancel)
Note that “⇒” indicates that there is a strong semantic relationship between words, and the strong semantic relationship here is a word that appears in the same phrase (an independent word) or a dependency relationship. It means a word (independent word) that appears in a phrase pair. For example, “information ⇒ search” means that “information” and “search” appear in the same phrase, or “information” and “search” appear in a phrase pair that has a dependency relationship, as shown below. Represents.
[0036]
  Sentence ： "Information search"
  Dependent clause pair: “Information” → “Search”
Also, the direction of “⇒” represents the order of appearance of words, and if the order of appearance is reversed, the meaning may be different, so the word order is important.
[0037]
  And in the concept expression, it is possible to connect and express any number of words (independent words). For example, in the above example 3), three words are connected. In this case, the three words appear in a strong relationship continuously in the word order of “information” “search” “service”. is doing. Therefore, A to D shown below conform to the conceptual expression of 3) in the above example, but E and F do not conform.
[0038]
  A: "Information search service"
  B: "Information search service"
  C: "Information search service"
  D: "Information search service"
  E: “Service that automatically searches for information”
  F: “Search Information Service”
The above example 4) shows an example in which a semantic tag is added, and the semantic tag uses the meaning as a tag when an expression such as an attached word in the phrase represents a specific additional meaning. Is. For example, the meaning tags “cancellation”, “request”, “possible”, and “question” are added to the phrase when the following words appear in the phrase. A plurality of semantic tags may be attached to one phrase.
[0039]
  Cancellation: Auxiliary verb "None", Auxiliary verb "Zu", Auxiliary verb "Mai", Auxiliary auxiliary verb "Hard", Adjective "None"
  Request: auxiliary verb "tai", verb "wanted", connecting particle "te" + verb "wanted"
  Question: Final particle "ka", final particle "ka" + final particle "na", symbol "?"
  Possible: auxiliary verb “can”, auxiliary verb “re”, auxiliary verb “re”
In the conceptual expression, for example, the meaning tag is represented by an expression such as “(+ cancellation + possible)”, and the meaning tag can be a concept expression alone or “execution (+ possible + cancellation)”. It can also be used in the form of words. For example, a phrase “unexecutable” is divided into “executable / can / cannot”, and a semantic tag “(+ possible + cancellation)” is added to this clause. The conceptual expression “execution (+ possible + cancellation)” means that the word “execution” is a phrase to which the semantic tags “cancellation” and “possibility” are added.
[0040]
  By using such a concept expression, the user can express a concept expression according to the purpose by a combination of an arbitrary number of words (independent words) and a semantic tag.
[0041]
  Then, for example, when “FAX → reception (+ cancellation)” is designated as the concept expression, the document processing apparatus 1 performs the concept extraction process in the concept extraction unit 6 as shown in FIG. .
[0042]
  First, the document processing apparatus 1 starts with a phrase with a phrase ID: k = 1 within a sentence ID: s = 1 within a document with a document ID: d = 1 in the document data structure (step S101). A suitability determination process is performed to check whether the structure after the clause k matches the designated concept expression (step S102). In the case of the above example, the following structure is applicable.
[0043]
  1) The phrase k includes the words “FAX” and “receive” in this order, and the intention tag “cancel” is added to the phrase k.
[0044]
  2) The phrase “k” includes the word “FAX”, the receiving clause k ′ in which the clause k is a related clause includes the word “reception”, and the intention tag “cancel” is added to the receiving clause k ′.
[0045]
  If the document ID does not match in step S102, the document processing apparatus 1 increments the phrase ID k in the sentence ID: s = 1 (k = k + 1), moves to the next phrase, and returns to the matching determination process (step S102). ).
[0046]
  In step S102, the concept expression extraction unit 6 searches for a word (independent word) that has a strong relationship in front of the concept expression specified by the concept expression specifying unit 7 when it is matched. It is registered in the expression extraction result list (step S103).
[0047]
  For example, in the case of the above example, the following words are suitable as the independent words.
[0048]
  1) The phrase k includes the word (independent word) X before the word “FAX”.
[0049]
  2) There is a related clause k 'in which the clause k is a receiving clause, and a word (independent word) X exists in the clause k'.
[0050]
  When such a word (independent word) is found, the concept expression extraction unit 6 registers it in the concept expression extraction result list. In this concept expression extraction result list, a word newly found in the designated concept expression is registered. Register the added new conceptual expression. For example, when the word “color” is newly found for “FAX” in the above example, the following concept expressions are registered in the concept expression extraction result list.
[0051]
  Color ⇒ FAX ⇒ Receive (+ Cancel)
In the concept expression extraction result list, as shown in FIG. 5, the total appearance frequency and the number of appearing documents of the registered concept expressions are also registered. At this time, the concept expression extraction unit 6 adds “1” to the appearance frequency when the concept expression to be registered has already been registered, and adds “1” to the number of appearing documents when registration from the document being processed is the first time. Add 1 ”. Further, when the concept expression is registered in the list for the first time, the concept expression extracting unit 6 sets the appearance frequency and the number of appearing documents to “1”.
[0052]
  Next, the concept expression extraction unit 6 searches for a word (independent word) having a strong relationship behind the concept expression specified by the concept expression specifying unit 7, and when an independent word is found, registers it in the concept expression extraction list ( Step S104).
[0053]
  For example, in the case of the above example, the following words are suitable as the independent words.
[0054]
  3) The phrase k includes the word (independent word) X after the word “FAX”.
[0055]
  4) There is a receiving clause k 'in which the clause k is a related clause, and a word (independent word) X exists in the clause k'.
[0056]
  When such a word (independent word) is found, the concept expression extraction unit 6 registers it in the concept expression extraction result list. In this concept expression extraction result list, a word newly found in the designated concept expression is registered. Register the added new conceptual expression. For example, when the word “symptom” is newly found for “FAX” in the above example, the following concept expressions are registered in the concept expression extraction result list.
[0057]
  FAX-> reception (+ cancellation)-> symptoms
As shown in FIG. 5, the concept expression extraction unit 6 also registers the total appearance frequency and the number of appearance documents of the registered concept expression in the concept expression extraction result list. At this time, the concept expression extraction unit 6 adds “1” to the appearance frequency when the concept expression to be registered has already been registered, and adds “1” to the number of appearing documents when registration from the document being processed is the first time. Add 1 ”. Further, when registering for the first time in the list, the concept expression extracting unit 6 sets the appearance frequency and the number of appearing documents to “1”.
[0058]
  Next, the concept expression extraction unit 6 checks whether a semantic tag that does not match the specified concept expression is added to the last clause of the document data structure that matches the specified concept expression. If a meaning tag is found, it is registered in the concept expression extraction result list (step S105).
[0059]
  As such a semantic tag, for example, in the case of the concept expression “FAX⇒receive (+ cancellation)” in the above example, the concept expression extraction unit 6 uses the rearmost document data structure that matches the designated concept expression. If the phrase k ′ has a meaning tag other than “cancel”, the meaning tag X is extracted. When such a semantic tag is found, the conceptual expression extraction unit 6 registers it in the conceptual expression extraction result list, but adds a newly found semantic tag to the specified conceptual expression. Register new conceptual expressions. For example, in the above example, when a new semantic tag “possible” is found, the following concept expressions are registered in the concept expression extraction result list.
[0060]
  FAX⇒Receive (+ Cancel + Possible)
Further, when a plurality of semantic tags are extracted, the concept expression extraction unit 6 registers a plurality of concept expressions in a form to which each is added. In the concept expression extraction result list, the total appearance frequency and the number of appearing documents of the registered concept expressions are also registered. At this time, the concept expression extraction unit 6 adds “1” to the appearance frequency when the concept expression to be registered has already been registered, and adds “1” to the number of appearing documents when registration from the document being processed is the first time. Add 1 ”. Further, when the concept expression is registered in the list for the first time, the concept expression extracting unit 6 sets the appearance frequency and the number of appearing documents to “1”.
[0061]
  Next, the conceptual expression extraction unit 6 checks whether the phrase ID: k (sentence k) is the last phrase in the sentence ID: s (sentence s) (step S106), and the phrase k is the last in the sentence s. If it is not, the phrase k is incremented by “1” (k = k + 1), the process returns to step S102, and the next phrase is processed in the same manner as described above from the matching determination process (steps S102 to S106).
[0062]
  When the phrase k is the last phrase in the sentence s in step S106, the conceptual expression extraction unit 6 determines whether the sentence ID: s (sentence s) is the last sentence in the document ID: d (document d). When the check is made (step S107) and the sentence s is not the last sentence in the document d, the sentence s is incremented by “1” (s = s + 1), the process returns to step S102, and the above-mentioned conformity determination is made for the next sentence. From the processing, the same processing as above is performed (steps S102 to S107).
[0063]
  When the sentence s is the last sentence in the document d in step S107, the conceptual expression extraction unit 6 checks whether the document ID: d (document d) is the last document (step S108), and the last sentence. If not, the sentence d is incremented by “1” (d = d + 1), the process returns to step S102, and the next phrase is processed in the same manner as described above from the matching determination process (steps S102 to S108).
[0064]
  If the document d is the last document in step S108, the concept expression extraction unit 6 determines that the concept extraction process has been completed for all the documents, and ends the process.
[0065]
  Then, the document processing apparatus 1 displays the result extracted by the concept expression extraction unit 6 on the concept expression display unit 8 as shown in FIGS. First, as shown in FIGS. 7 and 8, the document processing apparatus 1 first displays the word list (see FIG. 5) generated by the document data structure generating unit 4 on the left side, and the user displays the conceptual expression specifying unit 8. The result of the concept expression extraction using the word specified in the above as the specified concept expression is displayed on the right side. Further, in the display of the concept extraction result, two types of narrowing down by word shown in FIG. 7 and narrowing down by intention tag shown in FIG. 8 are selected and displayed. Further, in the document processing apparatus 1, when a user selects a concept expression displayed as a concept expression extraction result on the concept expression display unit 8, the concept expression is further extracted using the selected concept expression as a designated concept expression. You can also. In this way, conceptual expression extraction can be repeated.
[0066]
  Further, as shown in FIGS. 9 and 10, the document processing apparatus 1 first displays the intention tag list on the left side, and extracts the intention tag designated by the user in the concept expression designation unit 8 as the designated intention tag. The result of performing is displayed on the right side. Further, in the display of the extraction result, two types of narrowing down by word shown in FIG. 9 and narrowing down by intention tag shown in FIG. 10 are selected and displayed. Further, in the document processing apparatus 1, when a user selects a concept expression displayed as a concept expression extraction result on the concept expression display unit 8, the concept expression is further extracted using the selected concept expression as a designated concept expression. You can also. In this way, conceptual expression extraction can be repeated.
[0067]
  Further, as shown in FIGS. 11 and 12, the document processing apparatus 1 stores conceptual expressions designated by the user and displays a list thereof. Also in this case, the word as shown in FIG. 11 and the intention tag as shown in FIG. 12 are displayed.
[0068]
  In addition, in the display of FIGS. 7 to 12, it is possible to receive only a notation input from the user and display only a conceptual expression including the notation, whereby the user can see only necessary information.
[0069]
  Further, as shown in FIG. 14, the document processing apparatus 1 displays a document list including the designated conceptual expression. That is, the document processing apparatus 1 records a document including the concept expression specified by the concept expression specifying unit 7 in the concept expression extracting unit 6 and displays the document on the concept expression display unit 8 based on the stored document.
[0070]
  As described above, the document processing apparatus 1 according to the present embodiment performs morphological analysis and dependency analysis on the document (document set) input from the document input unit 2 and generates document data structure. The unit 4 converts the document into a document data structure holding linguistic information based on the language analysis result in the language analysis unit 3, stores the generated document data structure in the document data structure storage unit 5, and When designation input of an arbitrary concept expression is performed in the expression specifying unit 7, the concept expression specified by the concept expression specifying unit 7 from the document data structure stored in the document data structure storage unit 5 by the concept expression extracting unit 6. And the concept expression extraction result is displayed on the concept expression display unit 8.
[0071]
  Therefore, the user can freely search the concept expression included in the document to find the necessary concept expression which is not characteristic but can improve the usability.
[0072]
  Further, the document processing apparatus 1 of the present embodiment performs morphological analysis and dependency analysis on the document input from the document input unit 2 by the language analysis unit 3, and the document data structure generation unit 4 Is converted into a document data structure holding linguistic information based on the language analysis result in the language analysis unit 3, and the document data structure generated by the document data structure generation unit 4 is stored in the document data structure storage unit 5. Based on the document data structure, the concept representation to be selected and displayed is displayed on the concept representation display unit 8, and the concept representation designation unit 7 converts any concept representation from the concept display displayed on the concept representation display unit 8 When the selection is designated, the concept representation extraction unit 6 extracts a concept representation having a strong relationship with the concept representation designated by the concept representation designation unit from the document data structure stored in the document data structure storage unit 5. , Concept expression table Displaying the conceptual expression extraction result to part 8.
[0073]
  Therefore, it is possible to appropriately select from the displayed conceptual expressions, and the user can freely search through the conceptual expressions included in the document and easily and easily find the necessary conceptual expressions that are not characteristic. It can be discovered in a while, and the usability can be further improved.
[0074]
  Furthermore, in the document processing apparatus 1 according to the present embodiment, the document data structure generation unit 4 generates the document data structure by adding representative notation information to synonym / notation words based on the synonym dictionary. .
[0075]
  Therefore, it is possible to handle words having the same meaning but different notations as the same word, and the user can search the concept expression included in the document more freely and is not characteristic but necessary. It is possible to more easily find the concept expression to be performed, and the usability can be further improved.
[0076]
  Further, in the document processing apparatus 1 according to the present embodiment, the document data structure generation unit 4 adds a semantic tag representing an additional meaning to a word or phrase in the document data structure from an adjunct expression in the phrase. It is used as designation of the concept expression in the concept expression designation unit 7, extraction of the concept expression in the concept expression extraction unit 6, and display of the concept expression in the concept expression display unit 8.
[0077]
  Therefore, it is possible to specify a more detailed meaning, and to more appropriately find a necessary concept expression, thereby further improving usability.
[0078]
  Furthermore, the document processing apparatus 1 according to the present embodiment holds a history of designation of concept expressions in the concept expression designation unit 7, and based on the history, extracts a result of concept expression extraction based on the concept expression designation performed in the past. Redisplayed in the concept expression display unit 8.
[0079]
  Therefore, it is possible to immediately return to the work state performed by the user in the past, and the usability can be further improved.
[0080]
  Further, the document processing apparatus 1 according to the present embodiment displays a document including the concept expression specified by the concept expression specifying unit 7 or a part of the document on the concept expression display unit 8.
[0081]
  Therefore, it is possible to know how the concept expression specified by the user appears in the actual document, and it is possible to deepen the understanding of the concept expression and further improve the usability. it can.
[0082]
  The invention made by the present inventor has been specifically described based on the preferred embodiments. However, the present invention is not limited to the above, and various modifications can be made without departing from the scope of the invention. Needless to say.
[0083]
【The invention's effect】
  According to the present invention, a user can freely search within a conceptual expression included in a document to find a conceptual expression that is not characteristic but is necessary, and can improve usability. In particular,Based on independent words and ancillary words in the phrase, Semantic tags that represent additional meaning for words or phrases in the document data structure,That is,"Cancellation", "request", "question", "possible", etc. are given and the meaning tag is used for specifying the concept expression, extracting the concept expression, displaying the concept expression, etc., so that more detailed meaning can be specified It is possible to improve the usability by making it possible to find the necessary concept expression more appropriately.
[Brief description of the drawings]
FIG. 1 is a block diagram of a main part of a document processing apparatus to which an embodiment of a document processing apparatus, a document processing method, and a recording medium of the present invention is applied.
FIG. 2 is a diagram illustrating an example of language analysis in a language analysis unit in FIG. 1;
FIG. 3 is a diagram showing an example of conversion of a document set into a data structure of each document by the document data structure generation unit in FIG. 1;
4 is a diagram showing an example of information on each component of each data structure in FIG. 3;
5 is a diagram showing an example of ID, part of speech, appearance frequency, number of appearance documents, and synonym representative notation given to a word list included in a document or document set generated by the document data structure generation unit in FIG. 1; .
FIG. 6 is a diagram illustrating an example of a screen displayed on a concept expression display unit when a user directly inputs and specifies a concept expression.
FIG. 7 is a diagram illustrating an example of a screen displayed on a concept expression display unit when a user selects and designates a concept expression by word and narrows down by word.
FIG. 8 is a diagram showing an example of a screen displayed on a concept expression display unit when a user selects and designates a concept expression with a word and narrows down with an intention tag.
FIG. 9 is a diagram showing an example of a screen displayed on the concept expression display unit when the user selects and specifies with an intention tag and narrows down with words.
FIG. 10 is a diagram illustrating an example of a screen displayed on a concept expression display unit when a user selects and designates with an intention tag and narrows down with the intention tag.
FIG. 11 is a diagram showing an example of a screen displayed on the conceptual expression display unit when a user selects and specifies a history and narrows down by word.
FIG. 12 is a diagram showing an example of a screen displayed on the concept expression display unit when a user selects and specifies a history and narrows down with an intention tag.
13 is a flowchart showing concept extraction processing by the document processing apparatus of FIG. 1;
FIG. 14 is a diagram showing an example of a display screen of a document list including a designated concept expression on a concept expression display unit.
[Explanation of symbols]
    1 Document processing device
    2 Document input part
    3 Language Analysis Department
    4 Document data structure generator
    5 Document data structure storage
    6 Conceptual expression extraction unit
    7 Conceptual expression designation part
    8 Conceptual expression designation part

Claims (6)

Language analysis means for performing morphological analysis and dependency analysis on input document data;
Based on the result of language analysis in the language analysis means, a document data structure that holds the language information of the document data is generated, and when a specific independent word or ancillary word appears in a phrase, the document data structure Document data structure generating means for giving a meaning tag indicating cancellation, request, question or possibility according to the specific independent word or appendage that has appeared ,
Document data structure storage means for storing the document data structure to which the semantic tag is generated generated by the document data structure generation means;
A concept expression specifying means for accepting a concept expression consisting of a combination of an arbitrary number of words specified by the user and a word or phrase semantic tag in the document data structure;
A concept expression extracting means for extracting a concept expression related to the concept expression received by the concept expression specifying means from the document data structure stored in the document data structure storage means;
A document processing apparatus comprising: Further comprising a concept expression display means for displaying a concept expression to be selected and specified based on the document data structure,
The concept expression designation means accepts designation of an arbitrary concept expression from the concept display displayed on the concept expression display means,
The document processing apparatus according to claim 1, wherein the concept expression display unit displays a concept expression extraction result obtained by the concept expression extraction unit. A synonym dictionary
3. The document data structure generating unit generates the document data structure by adding representative notation information to synonym / notation words based on the synonym dictionary. Document processing device.   Means for holding a history of concept expressions received by the concept expression specifying means, and redisplaying the result of concept expression extraction based on the concept expression specification performed in the past on the concept expression display means based on the history; The document processing apparatus according to claim 2 or 3, wherein   The said concept expression display means displays the said document data containing the concept expression received by the said concept expression designation | designated means, or a part of the said document data, The one of Claim 2 thru | or 4 characterized by the above-mentioned. Document processing device.   The concept representation extraction unit checks whether the document data structure stored in the document data structure storage unit matches a concept representation (hereinafter, designated concept representation) received by the concept representation designation unit. From the data structure, a word having a dependency relationship with a word in the designated concept expression is extracted, a meaning tag not in the specified concept expression is extracted, and a concept expression including the extracted word and the meaning tag is created. The document processing apparatus according to claim 1, wherein the document processing apparatus includes:

Images