JP5454871B2 - Dictionary evaluation support apparatus and program - Google Patents

Description

  The present invention relates to a dictionary evaluation support apparatus and program for supporting evaluation work of a concept dictionary used in a business document check or text mining system.

  In recent years, a dictionary creation support technique for adding a new expression to a concept dictionary used in a business document check or text mining system or verifying the appropriateness of an already registered expression is known. In this dictionary creation support technology, for example, an expression matching a preset condition is extracted from a corpus (example sentence collection), and the dictionary creation can be supported by using the extracted expression.

  In relation to such a technique, for example, it is determined whether or not a dictionary needs to be updated, and a technique that can present a list of candidates to be registered in the dictionary when there is a need for updating (hereinafter referred to as a first preceding technique). (Referred to as Patent Document 1). According to this prior art, the co-occurrence expression (dependency expression) extracted from the corpus is set as a new dictionary registration candidate.

  In addition, for example, a technology that can reduce the burden of creating and maintaining data, can automatically expand different notations, and can create different notation dictionaries with more vocabularies (hereinafter referred to as the second prior art). Notation) is disclosed (see, for example, Patent Document 2). According to the second prior art, different notation candidates are extracted from the corpus and evaluated. Specifically, a part of kanji is converted to hiragana or converted according to a regulation rule such as converting the katakana notation “-” to “a”, and the similarity in the corpus is evaluated. It is determined whether or not there is.

Japanese Patent No. 3764618 JP 2004-110633 A

  In the first prior art described above, for example, by extracting and presenting expressions that co-occur with the same word for the purpose of daily report analysis, a concept extraction dictionary for the purpose of classifying documents by contents, such as the daily report analysis Can help you create.

  However, in the first prior art, for the purpose of, for example, the unification of notation and the extraction of completely synonymous ambiguous expressions, such as a business document check dictionary, even expressions with similar concepts are extracted and presented. As a result, the precision is reduced.

  In the second prior art described above, a combination of synonymous / same kanji and hiragana can be extracted, but a combination of kanji notations (for example, “here” and “here”) is extracted. I can't.

  For this reason, there is a need for a mechanism for evaluating the appropriateness of expressions already registered in a concept dictionary used in, for example, a business document check or text mining system, and creating a dictionary free of expression registration.

  Accordingly, an object of the present invention is to provide a dictionary evaluation support apparatus and program capable of supporting a dictionary evaluation work such as a business document check dictionary.

According to one aspect of the present invention, an example sentence corpus that stores in advance a plurality of example sentences including independent words, a morphological analysis dictionary that stores in advance independent words, reading of the independent words, and parts of speech of the independent words, A pattern match dictionary that stores in advance a first pattern that represents an expression that includes the independent word, and an expression that is represented by the first pattern by analyzing the first pattern stored in the pattern match dictionary Extracting means for extracting a first independent word included in the morphological analysis dictionary and searching for a second independent word having the same reading and part of speech as the extracted first independent word from the morphological analysis dictionary And generating a second pattern representing an expression including the second independent word based on the first pattern stored in the pattern match dictionary and the searched second independent word. A first example sentence that includes an expression represented by the first pattern by applying a first pattern stored in the pattern match dictionary to the example sentence stored in the example sentence corpus by applying pattern generation means and the first pattern stored in the pattern match dictionary By applying the first pattern matching means for creating the pattern match result of the second pattern created by the pattern creating means to the example sentences stored in the example sentence corpus, the second pattern is represented by the second pattern. Second pattern matching means for creating a pattern match result of a second example sentence including the expression to be expressed, and the number of first example sentences including the expression represented by the first pattern stored in the pattern match dictionary Determining means for determining whether or not is equal to or greater than a predetermined number; and the number of the first example sentences is not equal to or greater than a predetermined number If it is a constant, add to the third example sentence containing a representation represented by the first pattern stored in the pattern matching dictionary searching externally corpus, adds the third sentence in the example sentence corpus And a pattern matching result of the first example sentence created by the first pattern matching means, a pattern matching result of the second example sentence created by the second pattern matching means, and the first pattern. And a result output means for outputting a pattern match result of the third example sentence created based on the expression and the third example sentence . A dictionary evaluation support apparatus is provided.

  The present invention makes it possible to support an evaluation operation of a dictionary such as a business document check dictionary.

The block diagram which shows the hardware constitutions of the dictionary evaluation assistance apparatus which concerns on the 1st Embodiment of this invention. The block diagram which mainly shows a function structure of the dictionary evaluation assistance apparatus 30 shown in FIG. The block diagram which shows the structure of the user terminal device 40 shown in FIG. The figure which shows an example of the data structure of the example sentence corpus 22 shown in FIG. The figure which shows an example of the data structure of the morphological analysis dictionary 23 shown in FIG. The figure which shows an example of the data structure of the pattern match dictionary 24 shown in FIG. The flowchart which shows the process sequence of the dictionary evaluation assistance apparatus 30 which concerns on this embodiment. The figure which shows an example of a pattern matching process result. The figure which shows an example of the pattern matching result produced by the pattern matching part 35. FIG. The flowchart which shows the process sequence of an example sentence addition process. The figure which shows an example of the example sentence searched by the example sentence search part 371. The figure which shows an example of the syntax analysis result of the example sentence shown in FIG. The figure which shows an example of the dictionary evaluation assistance information shown to a user. The figure which shows an example of the data structure of the morphological analysis dictionary 23 contained in the dictionary evaluation assistance apparatus 30 which concerns on the 2nd Embodiment of this invention. The figure which shows an example of the data structure of the morphological analysis dictionary 23 contained in the dictionary evaluation assistance apparatus 30 which concerns on the 3rd Embodiment of this invention. The figure which shows an example of the data structure of the morphological analysis dictionary 23 contained in the dictionary evaluation assistance apparatus 30 which concerns on the 4th Embodiment of this invention. The figure which shows an example of the selection screen displayed by the display part 43 contained in the user terminal device 40. FIG. The flowchart which shows the process sequence of the dictionary evaluation assistance apparatus 30 which concerns on the 5th Embodiment of this invention. The flowchart which shows the process sequence of an example sentence deletion process. The figure which shows an example of the example sentence containing the corresponding expression "/ meeting [noun]" of the target applicable pattern contained in the pattern matching result produced by the pattern matching part. The block diagram which mainly shows a function structure of the dictionary evaluation assistance apparatus 300 which concerns on the 6th Embodiment of this invention. The flowchart which shows the process sequence of the example sentence addition process in the dictionary evaluation assistance apparatus 300 which concerns on this embodiment.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

[First Embodiment]
FIG. 1 is a block diagram showing a hardware configuration of the dictionary evaluation support apparatus according to the first embodiment of the present invention.

  As shown in FIG. 1, the computer 10 is connected to an external storage device 20 such as a hard disk drive (HDD). The external storage device 20 stores a program 21 executed by the computer 10. The computer 10 and the external storage device 20 constitute a dictionary evaluation support device 30.

  Further, the dictionary evaluation support device 30 (the computer 10 constituting the dictionary evaluation support device 30) is connected to a user terminal device 40 that uses the dictionary evaluation support device 30 via, for example, the network 50.

  FIG. 2 is a block diagram mainly showing a functional configuration of the dictionary evaluation support device 30 shown in FIG. As shown in FIG. 2, the dictionary evaluation support apparatus 30 includes a morpheme analysis unit 31, a regular expression analysis unit 32, a morpheme analysis dictionary search unit 33, a dictionary pattern creation unit 34, a pattern matching unit 35, a result determination unit 36, and an example sentence addition Part 37 and result output part 38. In the present embodiment, these units 31 to 38 are realized by the computer 10 illustrated in FIG. 1 executing the program 21 stored in the external storage device 20. This program 21 can be stored in advance in a computer-readable storage medium and distributed. Further, the program 21 may be downloaded to the computer 10 via the network 50, for example.

  The dictionary evaluation support device 30 includes an example sentence corpus 22, a morpheme analysis dictionary 23, and a pattern match dictionary 24. In the present embodiment, these units 22 to 24 are stored in, for example, the external storage device 20 (or a nonvolatile memory device).

  The example sentence corpus 22 stores (registers) a plurality of example sentences in advance. The plurality of example sentences stored in the example sentence corpus 22 include independent words such as nouns or verbs, for example.

  The morpheme analysis dictionary 23 is used for morphological analysis of sentences such as example sentences stored in the example sentence corpus 22, for example. In the morphological analysis dictionary 23, an independent word, a reading of the independent word, and a part of speech of the independent word are stored in advance.

  In the pattern match dictionary 24, a corresponding pattern (first pattern) representing an expression to be extracted from a sentence such as an example sentence stored in the example sentence corpus 22 is stored in advance. The corresponding pattern stored in the pattern match dictionary 24 includes an independent word (first independent word). Further, the corresponding pattern stored in the pattern match dictionary 24 is described by a regular expression, for example. The pattern match dictionary 24 is a concept dictionary used for, for example, business document check, and is an evaluation target in the dictionary evaluation support device 30.

  The morpheme analysis unit 31 performs a morpheme analysis process on the example sentences stored in the example sentence corpus 22. The morpheme analysis unit 31 executes morpheme analysis processing using the morpheme analysis dictionary 23. The morpheme analysis unit 31 stores the morpheme analysis result for the example sentence stored in the example sentence corpus 22 in the example sentence corpus 22 in association with the example sentence.

The morpheme analysis process is a process for dividing a sentence into words and adding parts of speech to the words. For example, if the morphological analysis process is executed for “Meet with customer”, “/ Customer [noun]-and [Participant] / Meeting [noun]-[Participant] / [Verb] /. [Phrase]] The result of morphological analysis is obtained. In the morpheme analysis results, “/” indicates a segment break, “-” indicates a word break in the clause, and the expression enclosed in “[]” indicates the part of speech of the immediately preceding word. The morphological analysis process for the example sentence by 31 may be executed in accordance with, for example, an instruction of a user who operates the user terminal device 40 or may be executed each time an example sentence is added to the example sentence corpus 22.

  The regular expression analysis unit 32 analyzes a corresponding pattern (hereinafter referred to as a target corresponding pattern) selected (designated) by the user via the user terminal device 40 among the corresponding patterns stored in the pattern match dictionary 24. . Thereby, the regular expression analysis part 32 extracts the independent word contained in the expression represented by the target corresponding pattern from the target corresponding pattern.

  The morpheme analysis dictionary search unit 33 is an independent word (second independent word) having the same sound and part of speech (the same reading and part of speech stored in the morpheme analysis dictionary 23) as the independent word extracted by the regular expression analysis unit 32. Is retrieved from the morphological analysis dictionary 23.

  The dictionary pattern creation unit 34 creates a corresponding pattern (second pattern) representing an expression including the independent word to be extracted from the sentence, based on the target corresponding pattern and the independent word searched by the morphological analysis dictionary searching unit 33. To do. In this case, the dictionary pattern creating unit 34 uses the independent words searched by the morphological analysis dictionary searching unit 33 as regular expressions of the target corresponding patterns stored in the pattern match dictionary 24 (regular expressions described as the target corresponding patterns). ) Is added (inserted) to create the corresponding pattern. Hereinafter, the corresponding pattern created by the dictionary pattern creating unit 34 is referred to as a pattern matching corresponding pattern.

  The pattern matching unit 35 applies the target corresponding pattern and the pattern matching corresponding pattern to the example sentence stored in the example sentence corpus 22 and the morphological analysis result of the example sentence. Thereby, the pattern matching unit 35 matches the pattern matching result of the example sentence (first example sentence) that matches the target corresponding pattern (including the expression represented by the target corresponding pattern) and the corresponding pattern for pattern matching (the relevant A pattern match result of an example sentence (second example sentence) including an expression represented by the corresponding pattern for pattern matching is created.

  Based on the pattern match result created by the pattern matching unit 35, the result determination unit 36 counts the number of example sentences including the expression represented by the target corresponding pattern (example sentences matching the target corresponding pattern). It is determined whether the number of is greater than or equal to a predetermined number.

  The example sentence adding unit 37 executes a process of adding an example sentence (third example sentence) including the expression represented by the target corresponding pattern to the example sentence corpus 22. The example sentence adding unit 37 includes an example sentence searching unit 371 and an example sentence evaluating unit 372.

  The example sentence search unit 371 searches for an example sentence including an expression represented by the target pattern, for example, from an external corpus.

  The example sentence evaluation unit 372 parses the example sentence searched by the example sentence search unit 371. The example sentence evaluation unit 372 determines whether or not the example sentence searched by the example sentence search unit 371 is an appropriate Japanese sentence based on the syntax analysis result. At this time, the example sentence evaluation unit 372 executes determination processing based on, for example, a predetermined condition.

  An example sentence determined by the example sentence evaluation unit 372 as an appropriate Japanese sentence is added to the example sentence corpus 22. In addition, the example sentence evaluation unit 372 creates a pattern match result (pattern match result of the third example sentence) of the example sentence (example sentence including the expression represented by the target pattern) determined to be an appropriate Japanese sentence, This is added to the pattern matching result created by the pattern matching unit 35.

  The result output unit 38 uses the pattern matching result (the pattern matching result of the first example sentence, the pattern matching result of the second example sentence, and the pattern matching result of the third example sentence) created by the pattern matching unit 35 as dictionary evaluation support information. To the user terminal device 40. The pattern match result output by the result output unit 38 is presented to the user via the user terminal device 40.

  FIG. 3 is a block diagram showing a configuration of the user terminal device 40 shown in FIG. As described above, the user terminal device 40 is connected to the dictionary evaluation support device 30 via the network 50, for example. The user terminal device 40 is a terminal device handled by a user who performs dictionary evaluation work using the dictionary evaluation support device 30.

  As illustrated in FIG. 3, the user terminal device 40 includes a terminal control unit 41, a terminal storage unit 42, a display unit 43, an input unit 44, and a communication interface 45. The terminal control unit 41, the terminal storage unit 42, the display unit 43, the input unit 44, and the communication interface (I / F) 45 are connected to each other inside the user terminal device 40, for example, via a bus.

  The terminal control unit 41 manages the entire process of the user terminal device 40. The terminal storage unit 42 is a storage device configured by hardware such as a hard disk drive or a nonvolatile memory device. For example, a program for controlling the user terminal device 40 is stored in the terminal storage unit 42. The terminal storage unit 42 also functions as a work memory for various processes performed by the terminal control unit 41.

  The display unit 43 is, for example, a liquid crystal disk play. The display unit 43 displays (presents) the pattern match result (dictionary evaluation support information) output by the above-described dictionary evaluation support device 30 (result output unit 38 included therein) to the user. In this case, the display unit 43 displays the pattern match result to the user in a table format, for example.

  The input unit 44 is, for example, a keyboard or a mouse. The input unit 24 receives, for example, an operation for dictionary evaluation work by the dictionary evaluation support device 30 from the user.

  The communication interface 45 is connected to the dictionary evaluation support device 30 via, for example, a cable. The communication interface 45 exchanges conditions specified by the user in the user terminal device 40 or processing results of the dictionary evaluation support device 30. The communication interface 45 can be connected to the external storage device 20 of the dictionary evaluation support device 30 via a cable, and the example sentence corpus 22 and the morpheme stored in the external storage device 20 are connected to the external storage device 20. It is also possible to execute input / output processing for the analysis dictionary 23 and the pattern match dictionary 24.

  FIG. 4 shows an example of the data structure of the example sentence corpus 22 shown in FIG. As shown in FIG. 4, the example sentence corpus 22 stores an example sentence (a plain text) and a morpheme analysis result (morpheme analysis sentence) of the example sentence in association with a number (identification number) for managing the example sentence. The morpheme analysis result of the example sentence is a result of the morpheme analysis process for the example sentence by the morpheme analysis unit 31.

  In the example shown in FIG. 4, the example sentence corpus 22 is associated with the number “1” and the example sentence “Conducted with customer on the 8th” and the morphological analysis result “/ 8 [number] / day [noun] of the example sentence. ]-Contains [participant] / customer [noun] / first [noun]-and [participant] / meeting [noun] / enforcement [noun] /. [Phrase].

  In the example corpus 22, the example sentence “Today's meeting will be held in the meeting room from 15:00” in association with the number “2” and the morphological analysis result “/ today [noun]-[participant] / [Noun]-is [Participant] / 15 [Number] / Time [Noun]-to [Participant] / Meeting [Noun] / Room [Noun]-[Participant] / Perform [Verb] -Mas [Auxiliary Verb] / [[Punctuation]] is stored.

  The example corpus 22 is associated with the number “3” and the example sentence “Consult about future policy” and the morphological analysis result “/ future [noun]-[participant] / policy [noun]-[ [Participant] / [Verb] /. [Phrase]] "are stored.

  The example corpus 22 is associated with the number “105”, and the example sentence “Discussed the measures for handicap countermeasures and reported the results” and the morphological analysis result of the example sentence “/ failure [noun] / measure [noun]-[ Particle] / hand [noun]-> [participant] / typed [verb] / combined [verb] -te [participant] / result [noun]-> [particle] / report [noun] / do [verb] -ta [ Auxiliary verb] /. [Punctation] ”is stored.

  The example corpus 22 is associated with the number “106”, and the example sentence “Meeting with the member to confirm progress” and the morphological analysis result “/ member [noun]-and [particle] / progress [noun] of the example sentence. ] / Confirmation [Noun]-[Participant] / For [Noun]-[Participant] / Consultation [Noun]-[Participant] / and then [Verb] -Ta [Auxiliary verb] /. [Phrase] ing.

  Also, in the example sentence corpus 22, the example sentence “I will adjust the schedule of the meeting” in association with the number “107” and the morphological analysis result “/ meeting [noun]-[participant] / schedule [noun]” of the example sentence. / Adjust [noun]-[particle] / do [verb] -mas [auxiliary verb] /. [Phrase] "is stored.

  FIG. 5 shows an example of the data structure of the morphological analysis dictionary 23 shown in FIG. The morphological analysis dictionary 23 shown in FIG. 5 is shown in a table format, for example.

  As shown in FIG. 5, the morphological analysis dictionary 23 stores (registers) an independent word (word) entry word, a reading of the independent word, and a part of speech of the independent word in association with each other. In addition, when the independent word stored in the morphological analysis dictionary 23 is a predicate (for example, a verb), the morphological analysis dictionary 23 stores the utilization type of the independent word. In addition to the above information, for example, information such as information obtained by subdividing the part of speech or pronunciation information may be added to the morphological analysis dictionary 23.

  In the example shown in FIG. 5, the morphological analysis dictionary 23 stores the independent word headword “below”, reading “squid”, and part-of-speech “noun” in association with each other.

  In the morphological analysis dictionary 23, the headword “meeting” of independent words, the reading “housekeeping”, and the part of speech “noun” are stored in association with each other.

  In the morphological analysis dictionary 23, the headword “meeting” of independent words, the reading “housekeeping”, and the part of speech “noun” are stored in association with each other.

  In the morphological analysis dictionary 23, the headword “meeting”, the reading “house together”, and the part of speech “noun” are stored in association with each other.

  In the morphological analysis dictionary 23, the headword “meeting” of independent words, the reading “housekeeping”, and the part of speech “noun” are stored in association with each other.

  In the morphological analysis dictionary 23, the headword “meeting” of independent words, the reading “housekeeping”, and the part of speech “noun” are stored in association with each other.

  In the morphological analysis dictionary 23, the headword “shooting together”, the reading “house together”, and the part of speech “noun” are stored in association with each other.

  In the morphological analysis dictionary 23, the headword “shoot together”, the reading “house together”, and the part of speech “noun” are stored in association with each other.

  In the morphological analysis dictionary 23, the headword “debate together”, the reading “house together”, and the part of speech “noun” are stored in association with each other.

  In the morpheme analysis dictionary 23, the headword “debate together”, the reading “house together” and the part of speech “noun” are stored in association with each other.

  In the morphological analysis dictionary 23, the headword “meeting”, the reading “meeting” and the part of speech “noun” are stored in association with each other.

  The morphological analysis dictionary 23 stores the independent word “room”, the reading “room”, and the part of speech “noun” in association with each other.

  In addition, the morphological analysis dictionary 23 stores an independent word heading “meeting”, reading “house together”, part-of-speech “verb”, and utilization type “first-line utilization”.

  FIG. 6 shows an example of the data structure of the pattern match dictionary 24 shown in FIG. The pattern match dictionary 24 shown in FIG. 6 is shown in a table format, for example.

  As shown in FIG. 6, the pattern match dictionary 24 stores codes, group names, item names, corresponding patterns, exclusion patterns, correction candidates, and messages for users in association with each other. Hereinafter, a code, a group name, an item name, a corresponding pattern, an exclusion pattern, a correction candidate, and a message for a user (registered line) associated with each other in the pattern match dictionary 24 illustrated in FIG. 6 are referred to as pattern information.

  In the code (column), an identification number for managing pattern information (code, group name, item name, corresponding pattern, exclusion pattern, correction candidate, and message for user) stored in the pattern match dictionary 24 is described. The

  In the group name (column), information given by the user regarding the pattern information is described. When the pattern match dictionary 24 is a dictionary used for business document check to extract (examine) expressions that cause misunderstanding or misunderstanding in business documents, for example, the pattern information is used for the group name from any viewpoint. Is described.

  In the item name (column), a headline relating to the expression actually extracted from the pattern information (corresponding pattern) is described. This item name is given by the user, for example. When the pattern information is used, for example, to unify the notation fluctuation of “meeting”, “meeting” is described in the item name.

  The corresponding pattern represents an expression to be extracted from the sentence as described above. In the corresponding pattern (column), the morphological analysis result of the expression to be extracted from the sentence is described in a regular expression (pattern).

  A regular expression is a method of expressing a character string using a specific character (metacharacter). In regular expressions, for example, “^” is the beginning of the line, “$” is the end of the line, “.” Is any single character, “(|)” is one of the expressions separated by “|”, “[ ] "Selects one of the characters arranged inside, and" \ "means that the next character is not interpreted as a metacharacter.

  As described above, by describing the corresponding pattern with the regular expression of the morphological analysis result, for example, it is possible to distinguish between “I” at the end of the sentence and “I” in the sentence “I will investigate and answer immediately”. It becomes possible to do. Thereby, for example, the possibility of erroneous extraction of the corresponding expression to be extracted in processing such as business document check can be reduced.

  The exclusion pattern is an expression represented by the corresponding pattern (an expression that matches the corresponding pattern), but represents an expression that is not to be extracted. In the exclusion pattern (column), the morpheme analysis result is described in a regular expression in the same manner as the corresponding pattern described above.

  For example, if you want to unify “meeting” to “meeting” but not “meeting room”, which is a proper noun as the name of the room, the notation for unifying the notation is, for example, the morphological analysis result of “meeting room” is a regular expression Is described in the exclusion pattern string.

  In another example, since the range of “below” in the sentence “determined as follows” is not clear, it is assumed that the expression “below” is to be extracted as an ambiguous expression to be noted. In this case, for example, when an expression having a different meaning of “below” is not to be extracted, such as “set a number of 3 or less”, for example, there is a number expression before the expression of “below” The result of the morphological analysis of the expression is described in the exclusion pattern string as a regular expression. Further, for example, a morpheme analysis result of an expression in which the particle “no” follows the expression “below”, followed by a numerical expression is described in the exclusion pattern string as a regular expression.

  The correction candidates (columns) include expressions that are represented by the corresponding pattern (expressions that match the corresponding pattern) but not expressions that are represented by the excluded pattern (expressions that match the excluded pattern), that is, expressions to be extracted In this case, correction candidates for the expression to be presented to the user when is detected are described.

  In the message for users (column), a message presented to the user when an expression to be extracted is found is described.

  In the example shown in FIG. 6, pattern information 241 and 242 are registered (stored) in the pattern match dictionary 24.

  The pattern information 241 of the pattern match dictionary 24 includes a code “a001”, a group name “unification”, an item name “meeting”, a corresponding pattern “/ (meeting | meeting) \ [noun \]”, an exclusion pattern “/ meeting” \ [Noun \] / room \ [noun \] ", correction candidate" meeting "and user message" notation "meeting". Is included.

  The code “a001” is an identification number for managing the pattern information 241 of the pattern match dictionary 24.

  The group name “unification of notation” indicates that, for example, the pattern information 241 is used from the viewpoint of unification of notation in, for example, business document check.

  The item name “meeting” indicates a heading regarding an expression actually extracted from the pattern information 241. This item name “meeting” is given by the user, for example.

  The corresponding pattern “/ (meeting | meeting) \ [noun \]” represents the expressions “/ meeting [noun]” and “/ meeting [noun]” to be extracted. In this way, since the corresponding pattern is described with a regular expression, it can represent a plurality of expressions. The same applies to the exclusion pattern or other patterns.

  The exclusion pattern “/ meeting \ [noun \] / room \ [noun \]” is an expression that is represented by the corresponding pattern “/ (meeting | meeting) \ [noun \]” but is not subject to extraction “/ "Meeting [noun] / room [noun]".

  According to the corresponding pattern “/ (meeting | meeting) \ [noun \]” and the exclusion pattern “/ meeting \ [noun \] / room \ [noun \]”, “/ meeting [noun]” and “/ meeting” "[Noun]" is extracted, but "/ meeting [noun] / room [noun]" is not extracted.

  The correction candidate “meeting” indicates that the correction candidate of the expression “/ meeting [noun]” or “/ meeting [noun]” extracted by the corresponding pattern is “meeting”.

  Please unify the message “Notation“ Meeting ”for users. "Indicates a message presented to the user when the expression" / meeting [noun] "or" / meeting [noun] "is extracted by the corresponding pattern.

  The pattern information 242 of the pattern match dictionary 24 includes a code “b001”, a group name “ambiguous expression”, an item name “below”, a corresponding pattern “/ below \ [noun \]”, an exclusion pattern “(/ [0 -9] + \ [numbers \] / below \ [nouns \] | / below \ [nouns \]-\ [particles] / [0-9] + \ [numbers \]) " This is an ambiguous expression, so please limit the scope as shown in the following three items. Is included.

  The code “b001” is an identification number for managing the pattern information 241 of the pattern match dictionary 24.

  The group name “ambiguous expression” indicates that, for example, the pattern information 242 is used to extract an ambiguous expression in, for example, a business document check.

  The item name “below” is a heading regarding the expression actually extracted from the pattern information 242, and is given by the user, for example.

  The corresponding pattern “/ below \ [noun \]” represents the expression “/ below [noun]” to be extracted. Exclusion pattern "(/ [0-9] + \ [numeric \] / below \ [noun \] | / below \ [noun \]-\\ particle \] / [0-9] + \ [number \] ) "Represents an expression that is represented by the corresponding pattern" / below \ [noun \] "but is not to be extracted. This exclusion pattern “(/ [0-9] + \ [numeric \] / below \ [noun \] | / below \ [noun \]-\\ participate \] / [0-9] + \ [number \ ]) "Indicates an expression in which a number precedes" below "and an expression in which a particle" no "is followed by" no "followed by a number.

  Message for users “This is an ambiguous expression. Please limit the scope to the following three items”. "Indicates a message presented to the user when the expression" / hereinafter [noun] "is extracted by the corresponding pattern.

  Of the information included in the pattern information registered in the pattern match dictionary 24 shown in FIG. 6, the corresponding pattern is indispensable information, but other information depends on the purpose or management method of using the pattern match dictionary. May be deleted as appropriate.

  Next, a processing procedure of the dictionary evaluation support device 30 according to the present embodiment will be described with reference to the flowchart of FIG.

  It is assumed that the example sentence corpus 22 stores a plurality of example sentences and morphological analysis results of the example sentences. The morpheme analysis result of the example sentence is a result of the morpheme analysis process for the example sentence by the morpheme analysis unit 31. Here, the example sentence corpus 22 will be described as having the data structure shown in FIG. 4 described above.

  The morphological analysis dictionary 23 will be described as having the data structure shown in FIG. Further, the pattern match dictionary 24 will be described as having the data structure shown in FIG.

  First, the user selects pattern information (a code included therein) stored in the pattern match dictionary 24 to be evaluated by the dictionary evaluation support device 30 by operating the input unit 44 included in the user terminal device 40. Here, it is assumed that the code “a001” included in the pattern information 241 shown in FIG. 6 is selected. The code “a001” selected by the user is transmitted (output) to the dictionary evaluation support device 30 via the communication interface 45 included in the user terminal device 40.

  The regular expression analysis unit 32 included in the dictionary evaluation support device 30 reads the pattern information 241 including the code “a001” transmitted by the user terminal device 40 from the pattern match dictionary 24 (step S1). As described above, the read pattern information 241 includes a code, a group name, an item name, a corresponding pattern, an exclusion pattern, a correction candidate, and a message for the user. Further, the corresponding pattern and the exclusion pattern are described by regular expressions including independent words.

  An independent word is a word that can constitute a phrase alone with respect to an attached word. For example, nouns, pronouns, verbs, adjectives, adjective verbs, adverbs, conjunctions, conjunctions, and emotional verbs correspond to independent words.

  Hereinafter, a corresponding pattern included in the pattern information 241 read by the regular expression analysis unit 32 is referred to as a target corresponding pattern. Further, the exclusion pattern included in the pattern information 241 read by the regular expression analysis unit 32 is referred to as a target exclusion pattern.

  Next, the regular expression analysis unit 32 analyzes the target corresponding pattern and the target exclusion pattern (regular expressions thereof) included in the read pattern information 241 and determines the independent words (the regular expressions) included in the target corresponding pattern and the target exclusion pattern ( (Independent words in the regular expression) are extracted (step S2).

  Here, the regular expression analysis unit 32 starts the independent words “meeting ([noun])” and “meeting ([noun]”) from the target corresponding pattern “/ (meeting | meeting) \ [noun \]” included in the pattern information 241. ) ”Is extracted. Further, the regular expression analysis unit 32 uses the independent words “meeting ([noun])” and “room ([[noun \]”) from the exclusion pattern “/ meeting \ [noun \] / room \ [noun \]” included in the pattern information 241. Noun]) ”is extracted. Each part of speech of the independent words “meeting”, “meeting”, and “room” extracted by the regular expression analysis unit 32 is a noun.

  Here, although the regular expression analysis unit 32 has been described as extracting all independent words included in the target corresponding pattern and the target exclusion pattern, only the independent words whose part of speech is a noun are extracted from the independent words. It does not matter.

  The morpheme analysis dictionary search unit 33 searches the morpheme analysis dictionary 23 for words (independent words) that have the same sound and part of speech (the same reading and part of speech) as the independent words extracted by the regular expression analysis unit 32 (step S3). At this time, the morpheme analysis dictionary search unit 33 searches for independent words other than the independent words extracted by the regular expression analysis unit 32.

  According to the morpheme analysis dictionary 23 shown in FIG. 5, the morpheme analysis dictionary search unit 33 reads the independent word “meeting” extracted by the regular expression analysis unit 32, reads the same “house together”, and the part of speech is “noun”. As independent words that are "meeting", "meeting", "meeting", "meeting", "arguing", "arguing", "arguing" and "arguing" morphological analysis dictionary Search from 23.

  In the morphological analysis dictionary 23 shown in FIG. 5, the independent words “meeting” and “room” extracted by the regular expression analyzing unit 32 are not stored, and the independent words “same part of speech” are not stored. Autonomous words with the same sound and part of speech as “meeting” and “room” are not searched.

  Next, the dictionary pattern creation unit 34 creates a pattern matching pattern based on the independent words searched by the morpheme analysis dictionary search unit 33 (step S4). The pattern matching pattern includes a pattern matching pattern, a pattern matching exclusion pattern, and a plaintext pattern.

  The dictionary pattern creation unit 34 selects an independent word having the same sound and part of speech as the independent word extracted from the target corresponding pattern by the regular expression analyzing unit 32 among the independent words searched by the morphological analysis dictionary searching unit 33. The corresponding pattern for pattern matching is created by inserting it into the regular expression.

  The dictionary pattern creation unit 34 selects an independent word having the same sound and the same part of speech as the independent word extracted from the target exclusion pattern by the regular expression analysis unit 32 among the independent words searched by the morphological analysis dictionary search unit 33. A pattern matching exclusion pattern is created by inserting it into the regular expression.

  Further, the dictionary pattern creation unit 34 creates a plaintext pattern by removing a morphological analysis result symbol (for example, “[]” representing part of speech) from the target corresponding pattern (regular expression thereof). In other words, the plaintext pattern is a pattern that targets an expression that is represented by the target corresponding pattern but has a different part of speech.

  Here, the independent words “meeting”, “meeting”, “meeting”, “meeting”, “meeting” with the same sound and part of speech as the independent word “meeting” extracted from the target pattern by the regular expression analysis unit 32. , “Shooting”, “discussing”, and “discussing” are searched by the morphological analysis dictionary searching unit 33, the dictionary pattern creating unit 34 selects “/ (meeting | meeting | Meeting | Discussion | Shooting | Shooting | Discussion | Discussion) \ [Noun \] "is created. In other words, the corresponding pattern for pattern matching is the independent word “meeting”, “meeting”, “meeting”, “meeting”, “shooting”, “shooting”, “discussion” instead of the independent word “meeting”. And “discussion” is inserted into the regular expression of the target pattern.

  Note that independent words having the same sound and part of speech as the independent words “meeting” and “room” extracted from the exclusion pattern by the regular expression analyzing unit 32 are not searched by the morphological analysis dictionary searching unit 33. In this case, the pattern matching exclusion pattern is not created.

  Further, the dictionary pattern creation unit 34 creates “(meeting | meeting)” as a plaintext pattern obtained by removing the morphological analysis result symbol from the target pattern.

  Next, the pattern matching unit 35 performs pattern matching patterns created by the target corresponding pattern, the target exclusion pattern, and the dictionary pattern creation unit 34 with respect to the example sentence stored in the example sentence corpus 22 and the morphological analysis result of the example sentence. (Corresponding pattern for pattern matching, exclusion pattern for pattern matching, and plaintext pattern) are applied. Thereby, the pattern matching part 35 performs a pattern matching process (step S5). In the pattern matching process, an example sentence that matches each of the target corresponding pattern, the target excluded pattern, and the pattern matching pattern (that is, includes an expression represented by each of the target corresponding pattern, the target excluded pattern, and the pattern matching pattern). Extracted from the example sentence corpus 22 and a pattern match result of the example sentence is created. In the following description, the target corresponding pattern, the target exclusion pattern, and the pattern matching pattern may be simply referred to as a pattern.

  It is assumed that the patterns used in the pattern matching process have a higher priority in the order of a target exclusion pattern, a pattern matching exclusion pattern, a target corresponding pattern, a pattern matching corresponding pattern, and a plaintext pattern. Thus, for example, when the same part (expression) of one example sentence matches a plurality of patterns among the target corresponding pattern, the target excluded pattern, the pattern matching corresponding pattern, the pattern matching excluded pattern, and the plaintext pattern, the corresponding example sentence Are treated as matching high priority patterns. Hereinafter, a pattern having the highest priority among a plurality of patterns that match the same part of the example sentence is referred to as a highest priority pattern.

  In addition, when a plurality of locations (expressions) match a pattern in one example sentence, the plurality of locations are handled separately.

  The pattern matching result generated by the pattern matching unit 35 is, for example, an emphasis display such as coloring an example sentence (plain text) extracted from the example sentence corpus 22 or a part (expression) matching the pattern in the morphological analysis result of the example sentence. And temporarily saved.

  Here, FIG. 8 shows an example of a pattern matching process result to which the corresponding pattern, the exclusion pattern, and the pattern matching pattern (pattern matching corresponding pattern, pattern matching exclusion pattern, and plaintext pattern) are applied.

  In the example shown in FIG. 8, as an example sentence that matches the pattern, the example sentence managed by the number “1” “conducted with customer on the 8th” and the morphological analysis result “/ 8 [number] / day of the example sentence” [Noun]-[Participant] / Customer [Noun] / First [Noun]-and [Participant] / Association [Noun] / Execution [Noun] /. [Punct]]

  In the example shown in FIG. 8, as an example sentence matched with the corresponding pattern, an example sentence managed by the number “2” “Today's meeting will be held in the meeting room from 15:00” and the morphological analysis result “/ today [ Noun]-[Participant] / Meeting [Noun]-[Participant] / 15 [Number] / Time [Noun] -From [Participant] / Meeting [Noun] / Room [Noun]-[Participant] / Perform [Verb] -mass [auxiliary verb] /. [Phrase] "is shown.

  In the example shown in FIG. 8, as an example sentence that matches the exclusion pattern, an example sentence managed by the number “2” “Today's meeting will be held in the meeting room from 15:00” and the morphological analysis result “/ today [ Noun]-[Participant] / Meeting [Noun]-[Participant] / 15 [Number] / Time [Noun] -From [Participant] / Meeting [Noun] / Room [Noun]-[Participant] / Perform [Verb] -mass [auxiliary verb] /. [Phrase] "is shown.

  In the example shown in FIG. 8, as an example sentence matched with a plain text pattern, an example sentence managed by the number “3” “discusses about future policy” and a morphological analysis result “/ future [noun]-[participant] of the example sentence. ] / Policy [Noun]-[Participant] / Meet [Verb] /. [Phrase]].

  In the example shown in FIG. 8, as an example sentence matched with the corresponding pattern for pattern matching, an example sentence “meeting with a member to confirm progress” managed by the number “106” and a morphological analysis result “/ member” of the example sentence [Noun]-and [Participant] / Progress [Noun] / Confirmation [Noun]-[Participant] / For [Noun]-[Participant] / Association [Noun]-[Participant] / S [Verb] -Ta [Auxiliary verb] /. [Phrase] "is shown.

  In the example shown in FIG. 8, as an example sentence matched with the corresponding pattern for pattern matching, the example sentence “schedule adjustment is performed” managed by the number “107” and the morphological analysis result “/ meeting [ Noun]-[Participant] / Schedule [Noun] / Adjustment [Noun]-[Participant] / Perform [Verb] -Mas [Auxiliary Verb] /. [Puncture]] is shown.

  FIG. 9 shows an example of the pattern matching result created by the pattern matching unit 35. The pattern match results shown in FIG. 9 are created by sorting the pattern match processing results shown in FIG.

  Here, the corresponding expression refers to an individual expression (that is, each expression represented by the pattern) obtained by analyzing the regular expressions of the target corresponding pattern, the target exclusion pattern, and the pattern matching pattern and converting them into specific expressions. .

  The corresponding expressions of the target corresponding pattern “/ (meeting | meeting) \ [noun \]” are “/ meeting [noun]” and “/ meeting [noun]”.

  The corresponding expression of the exclusion pattern “/ meeting \ [noun \] / room \ [noun \]” is “/ meeting [noun] / room [noun]”.

  The corresponding expression for the pattern matching pattern "/ (Meeting | Meeting | Meeting | Meeting | Shooting | Shooting | Shooting | Discussing | Discussing) \ [Noun \]" is "/ Meeting [noun]", "/ "Meeting [noun]", "/ Meeting [noun]", "/ Meeting [noun]", "/ Meeting [noun]", "/ Meeting [noun]", "/ Meeting [noun]" and It is “/ debate [noun]”.

  The corresponding expressions of the plaintext pattern “(meeting | meeting)” are “meeting” and “meeting”.

  In the following description, a regular expression of a target corresponding pattern is analyzed, and each expression (that is, a corresponding expression) converted into a specific expression is simply referred to as a corresponding expression of the target corresponding pattern. The same applies to the target exclusion pattern and the pattern matching pattern.

  As shown in FIG. 9, the pattern matching result includes a matched top priority pattern, a corresponding expression, an example sentence, and a morphological analysis result of the example sentence. In the pattern match result, a number for managing the corresponding expression is assigned to each corresponding expression. In the following description, the matched top-priority pattern, corresponding expression, example sentence, and morphological analysis result of the example sentence included in the pattern match result may be referred to as information related to the example sentence for convenience.

  In the pattern matching result shown in FIG. 9, the corresponding expression of the matched top priority pattern is highlighted in the example sentence and the morphological analysis result of the example sentence.

  Further, in the pattern matching result shown in FIG. 9, when there is no example sentence matching the pattern (that is, an example sentence including the corresponding expression of the pattern) in the example sentence corpus 22, the pattern matching result is associated with the corresponding expression. No example sentence to be written ".

  Returning to FIG. 7 again, the result determining unit 36 and the example sentence adding unit 37 execute the example sentence adding process with reference to the pattern match result created by the pattern matching unit 35 (step S6). In this example sentence addition process, an example sentence including the relevant expressions of the target corresponding pattern and the target excluded pattern and information related to the example sentence are added to the pattern match result created by the example sentence corpus 22 and the pattern matching unit 35. Details of the example sentence addition process will be described later.

  The result output unit 38 outputs the pattern match result obtained by adding the example sentence (information related to the example sentence by the example sentence addition process) to the user terminal device 40 as dictionary evaluation support information (step S7).

  The pattern match result output by the result output unit 38 is input by the communication interface 45 of the user terminal device 40 and displayed on the display unit 43. Thereby, in the user terminal device 40, a pattern match result (dictionary evaluation support information) is presented to the user.

  Next, with reference to a flowchart of FIG. 10, a processing procedure of the above-described example sentence addition processing (processing in step S6 shown in FIG. 7) will be described. Here, description will be made assuming that the pattern matching result shown in FIG.

  The example sentence addition process is executed for each of the corresponding expressions of the target corresponding pattern and the target excluded pattern. Hereinafter, the corresponding expression that is the target of the example sentence addition process is referred to as a target corresponding expression.

  First, the result determination unit 36 refers to the pattern matching result created by the pattern matching unit 35 and counts (number of) example sentences including the target corresponding expression. That is, the result determination unit 36 counts the number of example sentences associated with the target corresponding expression in the pattern match result shown in FIG. 9 (hereinafter referred to as the number of example sentences of the target corresponding expression).

  More specifically, referring to FIG. 9, when the target corresponding expression is “/ meeting [noun]”, for example, the number of example sentences of the target corresponding expression “/ meeting [noun]” is two. Further, when the target corresponding expression is “/ meeting [noun]”, for example, the number of example sentences of the target corresponding expression “/ meeting [noun]” is zero. For example, when the target corresponding expression is “/ meeting [noun] / room [noun]”, the number of example sentences of the target corresponding expression “/ meeting [noun] / room [noun]” is one.

  Next, the result determination unit 36 determines whether or not the number of example sentences of the target expression is equal to or greater than a predetermined number (first specified number) (step S11). The first specified number indicates the lower limit of the number of example sentences of the target corresponding expression.

  For example, when the target corresponding expression is “/ meeting [noun]” and the first specified number is 1, the number of example sentences of the target corresponding expression is 0, so the number of example sentences of the target corresponding expression is the first. It is determined that the specified number is not exceeded.

  When it is determined that the number of example sentences of the target corresponding expression is not equal to or greater than the first specified number (NO in step S11), the example sentence searching unit 371 included in the example sentence adding unit 37 selects, for example, dictionary evaluation support for the example sentence including the target corresponding expression A search is performed from a corpus outside the device 30 (step S12). The example sentence search unit 371 searches for an example sentence whose morphological analysis result includes the target corresponding expression from a corpus such as a Web space.

  Here, FIG. 11 shows an example of an example sentence searched by the example sentence search unit 371 when the target corresponding expression is “/ meeting [noun]”. The example sentence searched by the example sentence search unit 371 is an example sentence including the target corresponding expression “/ meeting [noun]”, and the corresponding expression of the target exclusion pattern (here, “/ meeting [noun] / room [noun] ] ”) Is an example with no pattern match.

  As shown in FIG. 11, the example sentences searched by the example sentence search unit 371 include “Meeting (hereinafter referred to as example sentences of result_1)” and “Tomorrow, you will be meeting with your company. ”And“ The person in charge shows the customer and the meeting tomorrow. (Hereinafter, written as an example sentence of result3) ”.

  Next, the example sentence evaluation unit 372 parses the example sentence searched by the example sentence search unit 371 (step S13).

  Here, FIG. 12 shows an example of the syntax analysis result of the example sentence shown in FIG. In the parsing result shown in FIG. 12, for example, the dependency relation of clauses is shown in a branch structure with the predicate (eg “Ukaga” in the example sentence of result2 and “shimasu” in the example sentence in result3) as the root of the tree structure. Yes.

  The example sentence evaluation unit 372 determines whether there is an example sentence that is an appropriate Japanese sentence in the example sentence based on the syntax analysis result of the example sentence searched by the example sentence search unit 371 (step S14). The example sentence evaluation unit 371 determines, for example, that an example sentence that matches a predetermined condition (evaluation criteria) is an appropriate Japanese sentence.

  Here, the predetermined condition indicates, for example, that “predicate” and “subject” are included in the syntax analysis result. Note that this condition is an example, and other conditions may be used.

  According to the syntax analysis result of the example sentence shown in FIG. 12, the example sentence of result3 matches this condition. For this reason, the example sentence evaluation unit 372 determines that there is an example sentence that is an appropriate Japanese sentence.

  When the example sentence evaluation unit 372 determines that there is an example sentence that is an appropriate Japanese sentence (YES in step S14), the result determination unit 36 adds the example sentence to the example sentence corpus 22 and adds information on the example sentence to the pattern matching part. It adds to the pattern match result created by 35 (step S15).

  The information related to the example sentence added to the pattern matching result created by the pattern matching unit 35 includes the matched highest priority pattern, the corresponding expression, the example sentence, and the morphological analysis result of the example sentence. Note that the information about the example sentence (the pattern matching result of the third example sentence) is created by the example sentence evaluation unit 372 based on the target corresponding expression and the example sentence.

  When the process of step S15 is executed, it is determined whether or not the example sentence addition process (the process after step S11 described above) has been executed for all the corresponding expressions of the target corresponding pattern and the target excluded pattern (step S16).

  If it is determined that the example sentence addition process has not been executed for all the corresponding expressions (NO in step S16), the process returns to step S11 and the process is repeated. In this case, the process is executed with the corresponding expression that has not been processed as the target corresponding expression.

  On the other hand, when it is determined that the example sentence addition process has been executed for all the corresponding expressions (YES in step S16), the example sentence addition process is terminated.

  In this case, as described above, the pattern match result to which the information related to the example sentence (the example sentence retrieved by the example sentence retrieval unit 371 and determined to be an appropriate Japanese sentence by the example sentence evaluation unit 372) is added is the corresponding expression. Sorted by (target applicable pattern, target excluded pattern, and pattern matching pattern corresponding expression). The pattern match result is presented to the user via the user terminal device 40 as dictionary evaluation support information.

  Here, FIG. 13 shows an example of dictionary evaluation support information presented to the user. The dictionary evaluation support information illustrated in FIG. 13 indicates dictionary evaluation support information when information related to the example sentence of result3 illustrated in FIGS. 11 and 12 is added to the pattern match result illustrated in FIG. 9 described above.

  As shown in FIG. 13, the dictionary evaluation support information includes the matched top priority pattern, the corresponding expression, the example sentence, and the morphological analysis result of the example sentence, similarly to the pattern matching result shown in FIG.

  As shown in FIG. 13, the dictionary evaluation support information includes, in addition to the pattern match result, for example, the name of the pattern match dictionary 24 (pattern match dictionary name), the code selected by the user (the line of the pattern match dictionary 24). Code), corresponding pattern (target corresponding pattern), and exclusion pattern (target exclusion pattern).

  In the example shown in FIG. 13, the dictionary evaluation support information includes a pattern match dictionary name “pattern match dictionary 01”, a line code “a001”, a corresponding pattern “/ (meeting | meeting) \ [noun \]”, and an exclusion pattern. "/ Meeting \ [noun \] / room \ [noun \]" is included.

  As described above, in the dictionary evaluation support information presented to the user, the corresponding expression (corresponding expression represented by the corresponding pattern and the pattern for exclusion pattern pattern matching) in the example sentence including each corresponding expression and the morphological analysis result of the example sentence. Is highlighted.

  Thereby, the user refers to the above-described dictionary evaluation support information displayed by the user terminal device 40 (display unit 43 thereof), for example, by using the pattern information (corresponding pattern) registered in the pattern match dictionary 24. Since the extracted expression can be confirmed, the pattern matching dictionary 24 can be easily evaluated.

  If it is determined in step S11 shown in FIG. 10 that the number of example sentences is greater than or equal to the first specified number, the process of step S16 is executed. That is, since the number of example sentences including the target corresponding expression is sufficient, the addition process of the example sentence including the target corresponding expression is not performed.

  The example sentence addition process shown in FIG. 10 has been described as being executed for each of the corresponding expressions of the target corresponding pattern and the target excluded pattern. It does not matter.

  As described above, in the present embodiment, an independent word that has the same sound and the same part of speech as the independent word (first independent word) included in the expression represented by the corresponding pattern and the excluded pattern stored in the pattern match dictionary 24. The second independent word) is retrieved from the morphological analysis dictionary 23. In the present embodiment, pattern matching patterns (corresponding patterns for pattern matching, pattern matching exclusion patterns, and plaintext patterns) representing expressions including independent words retrieved from the morphological analysis dictionary 23 are created and stored in the pattern matching dictionary 24. The corresponding pattern, the exclusion pattern, and the pattern matching pattern stored are applied to the example sentence (and the morphological analysis result of the example sentence) stored in the example sentence corpus 22. In this embodiment, the pattern match result (matched top-priority pattern, corresponding expression, example sentence, and morphological analysis result of the example sentence) of the example sentence that matches the corresponding pattern, the exclusion pattern, and the pattern matching pattern is transmitted to the user as dictionary evaluation support information. To be presented.

  Thereby, in this embodiment, since the corresponding expression of the pattern for pattern matching included in the dictionary evaluation support information can be presented to the user as a dictionary registration addition candidate, the user refers to the dictionary evaluation support information. As a result, the registration omission of the ambiguous expression having the same meaning as the corresponding pattern (and the exclusion pattern) stored (registered) in the pattern match dictionary 24 (for example, “meeting” is registered while “meeting” is registered). And the like can be additionally registered in the pattern match dictionary 24.

  In the present embodiment, when the number of example sentences including the corresponding expression of the corresponding pattern or the exclusion pattern stored in the pattern match dictionary 24 is not equal to or greater than the prescribed number, the example sentence including the corresponding expression of the corresponding pattern or the exclusion pattern Is retrieved from an external corpus such as a web space. In the present embodiment, it is determined whether or not an example sentence retrieved from an external corpus is an appropriate Japanese sentence, and the example sentence determined to be the appropriate Japanese sentence is added to the example sentence corpus 22, and the example sentence Information about the pattern matching result (pattern matching result of the example sentence) is added to the pattern matching result (dictionary evaluation support information) created by the pattern matching unit 35.

  Thus, in the present embodiment, as described above, for example, even when an example sentence including the corresponding expression “/ meeting \ [noun \]” of the corresponding pattern is not stored in the example sentence corpus 22, dictionary evaluation is performed. An example sentence including the corresponding expression “/ meeting \ [noun \]” of the corresponding pattern can be added to the support information. For this reason, the user who is presented with the dictionary evaluation support information can confirm the example sentence including the corresponding expression “/ meeting \ [noun \]”, and therefore can appropriately evaluate the corresponding pattern (corresponding expression). it can.

  In the present embodiment, the example sentence and the morphological analysis result of the example sentence are sorted and output (presented) by the corresponding expression of the corresponding pattern, the exclusion pattern, and the pattern matching pattern. By referring to, whether the corresponding expression is correct or not can be evaluated efficiently.

  Furthermore, in this embodiment, by using a plaintext pattern as the pattern matching pattern, for example, the corresponding expression of the corresponding pattern registered in the pattern match dictionary 24 and the plaintext (sentence that has not been subjected to morphological analysis) are different. In this example, the user can present an example sentence including the same expression to the user, so that the user can find a missing dictionary registration of the corresponding expression having a different morphological analysis result and can additionally register it in the pattern match dictionary 24. Become.

  In the present embodiment, the pattern match dictionary 24 to be evaluated by the dictionary evaluation support apparatus 30 has been described as a concept dictionary used for business document check, for example. When a similar pattern match dictionary 24 is used in the processing, the pattern match dictionary 24 can be an evaluation target.

[Second Embodiment]
Next, a second embodiment of the present invention will be described. The functional configuration of the dictionary evaluation support apparatus according to this embodiment is the same as that of the first embodiment described above, and will be described with reference to FIG. 2 described above.

  This embodiment is different from the first embodiment described above in that the standard notation of independent words is further stored in the morphological analysis dictionary 23 included in the dictionary evaluation support device 30.

  Here, FIG. 14 shows an example of the data structure of the morphological analysis dictionary 23 included in the dictionary evaluation support device 30 according to the present embodiment.

  As shown in FIG. 14, the morphological analysis dictionary 23 stores the independence word when the independent word headword, the independent word reading shown in FIG. 5, the part of speech of the independent word, and the independent word are predicates. In addition to the word utilization type, the standard notation of the independent word is stored in association with it.

  In the example shown in FIG. 14, the morphological analysis dictionary 23 stores standard notation “below” in association with the headword “below”, for example.

  The morphological analysis dictionary 23 stores the standard title “meeting” in association with, for example, the headwords “meeting”, “meeting”, “meeting”, “meeting”, and “meeting”.

  In the morphological analysis dictionary 23, for example, the standard notation “shooting” is stored in association with the headwords “shooting” and “shooting”.

  In the morphological analysis dictionary 23, for example, the standard notation “discussion” is stored in association with the headwords “discussion” and “discussion”.

  In the morphological analysis dictionary 23, for example, the standard notation “meeting” is stored in association with the headword “meeting”.

  In the morphological analysis dictionary 23, for example, the standard notation “room” is stored in association with the headword “room”.

  The morphological analysis dictionary 23 stores a standard notation “meeting” in association with, for example, the headword “meeting”.

  Hereinafter, the operation of the dictionary evaluation support device 30 according to the present embodiment will be described. The morphological analysis dictionary search unit 33 included in the dictionary evaluation support device 30 according to the present embodiment is an independent word having the same standard notation as the independent word with the same sound and part of speech as the independent word extracted by the regular expression analyzing unit 32. (Independent words with the same standard notation) are searched from the morphological analysis dictionary 23.

  Note that the dictionary evaluation support apparatus 30 according to the present embodiment is the same as the dictionary evaluation support apparatus 30 according to the first embodiment described above except for the operation of the morphological analysis dictionary search unit 33 described above. Omitted.

  As described above, in the present embodiment, the morphological analysis dictionary search unit 33 searches for independent words that have the same sound and the same part of speech as the independent words extracted by the regular expression analysis unit 32 and that have the same standard notation. Finally, it is possible to omit expressions that have the same sound and part of speech but that are likely to have different meanings from the corresponding expression of the pattern matching pattern presented to the user as a dictionary registration addition candidate.

  Specifically, in the first embodiment described above, the corresponding expression of the corresponding pattern for pattern matching is “/ meeting [noun]”, “/ meeting [noun]”, “/ meeting [noun]”, “/ While this was a meeting [noun], "/ shoot together [noun]", "/ shoot together [noun]", "/ discuss [noun]" and "/ discuss [noun]" In the embodiment, the corresponding expression of the corresponding pattern for pattern matching is limited to “/ meeting [noun]”, “/ meeting [noun]”, “/ meeting [noun]”, “/ meeting [noun]”. Can do.

  Thereby, in this embodiment, the efficiency of the dictionary evaluation work based on dictionary evaluation support information can be improved.

[Third Embodiment]
Next, a third embodiment of the present invention will be described. The functional configuration of the dictionary evaluation support apparatus according to this embodiment is the same as that of the first embodiment described above, and will be described with reference to FIG. 2 described above.

  This embodiment is different from the first embodiment described above in that the meaning classification of independent words is further stored in the morphological analysis dictionary 23 included in the dictionary evaluation support device 30.

  Here, FIG. 15 shows an example of the data structure of the morphological analysis dictionary 23 included in the dictionary evaluation support device 30 according to the present embodiment.

  As shown in FIG. 15, the morphological analysis dictionary 23 includes the independent word headword shown in FIG. 5, the reading of the independent word, the part of speech of the independent word, and the independent word when the independent word is a predicate. In addition to the word utilization type, the meaning classification (information indicating) of the independent word is stored in association with each other.

  In the example shown in FIG. 15, the morphological analysis dictionary 23 stores, for example, a semantic classification “range specification” in association with the headword “below”.

  The morphological analysis dictionary 23 stores, for example, a semantic classification “meeting” in association with the headwords “meeting”, “meeting”, “meeting”, “meeting”, and “meeting”.

  The morphological analysis dictionary 23 stores, for example, the semantic classification “shooting action” in association with the headwords “shooting” and “shooting”.

  The morphological analysis dictionary 23 stores, for example, a semantic classification “attacking” in association with the headwords “discussion” and “discussion”.

  The morphological analysis dictionary 23 stores, for example, a semantic classification “meeting” in association with the headword “meeting”.

  The morphological analysis dictionary 23 stores, for example, a semantic classification “equipment” in association with the headword “room”.

  The morphological analysis dictionary 23 stores a semantic classification “meeting” in association with the headword “meeting”, for example.

  Hereinafter, the operation of the dictionary evaluation support device 30 according to the present embodiment will be described. The morphological analysis dictionary search unit 33 included in the dictionary evaluation support device 30 according to the present embodiment is an independent word having the same semantic classification as the independent word that has the same sound and part of speech as the independent word extracted by the regular expression analyzing unit 32. (Independent words with the same semantic classification) are searched from the morphological analysis dictionary 23.

  Note that the dictionary evaluation support apparatus 30 according to the present embodiment is the same as the dictionary evaluation support apparatus 30 according to the first embodiment described above except for the operation of the morphological analysis dictionary search unit 33 described above. Omitted.

  As described above, in the present embodiment, the morphological analysis dictionary search unit 33 searches for independent words that have the same sound and part of speech and the same semantic classification as the independent words extracted by the regular expression analysis unit 32. Finally, it is possible to omit expressions that have the same sound and part of speech but that are likely to have different meanings from the corresponding expression of the pattern matching pattern presented to the user as a dictionary registration addition candidate.

  Specifically, in the first embodiment described above, the corresponding expression of the corresponding pattern for pattern matching is “/ meeting [noun]”, “/ meeting [noun]”, “/ meeting [noun]”, “/ While this was a meeting [noun], "/ shoot together [noun]", "/ shoot together [noun]", "/ discuss [noun]" and "/ discuss [noun]" In the embodiment, the corresponding expression of the corresponding pattern for pattern matching is limited to “/ meeting [noun]”, “/ meeting [noun]”, “/ meeting [noun]”, “/ meeting [noun]”. Can do.

  Thereby, in this embodiment, the efficiency of the dictionary evaluation work based on dictionary evaluation support information can be improved.

[Fourth Embodiment]
Next, a fourth embodiment of the present invention will be described. The functional configuration of the dictionary evaluation support apparatus according to this embodiment is the same as that of the first embodiment described above, and will be described with reference to FIG. 2 described above.

  This embodiment is different from the first embodiment described above in that the meaning of an independent word is further stored in the morphological analysis dictionary 23 included in the dictionary evaluation support device 30.

  Here, FIG. 16 shows an example of the data structure of the morphological analysis dictionary 23 included in the dictionary evaluation support device 30 according to the present embodiment.

  As shown in FIG. 16, the morphological analysis dictionary 23 includes the above-mentioned independent word heading shown in FIG. 5, the reading of the independent word, the part of speech of the independent word, and the independent word when the independent word is a predicate. In addition to the utilization type, the meaning (information indicating) of the independent word is stored in association with each other.

  In the example shown in FIG. 16, the morphological analysis dictionary 23 stores, for example, the meaning “becomes a lower range in comparison of quantity, degree, superiority and inferiority” in association with the headword “squid”. .

  The morphological analysis dictionary 23 stores, for example, the meaning “consult in advance” in association with the headwords “meeting”, “meeting”, “meeting”, “meeting”, and “meeting”.

  The morphological analysis dictionary 23 stores, for example, the meaning “shooting” in association with the headwords “shooting” and “shooting”.

  The morphological analysis dictionary 23 stores, for example, the meaning “debating” in association with the headwords “debating” and “debating”.

  The morphological analysis dictionary 23 stores, for example, the meaning “small meeting” in association with the headword “meeting”.

  In the morphological analysis dictionary 23, for example, the meaning “room” is stored in association with the headword “room”.

  In addition, the morphological analysis dictionary 23 stores, for example, the meaning “consult in advance” in association with the headword “meeting”.

  Hereinafter, the operation of the dictionary evaluation support device 30 according to the present embodiment will be described. The morphological analysis dictionary search unit 33 included in the dictionary evaluation support device 30 according to the present embodiment is independent of the same words and parts of speech as the independent words extracted by the regular expression analysis unit 32, as in the first embodiment described above. The word is searched from the morphological analysis dictionary 23.

  The morpheme analysis dictionary search unit 33, after the search processing, information indicating the searched independent word, the part of speech of the independent word stored in the morphological analysis dictionary 23 in association with the independent word, and the meaning of the independent word ( ) To the user terminal device 40.

  The morphological analysis dictionary search unit 33 accepts a selection by the user as to whether or not the output independent word is to be the corresponding expression of the pattern matching pattern.

  In the user terminal device 40 (the display unit 43 included therein), a selection screen on which the independent words output by the morphological analysis dictionary search unit 33, the part of speech of the independent words, and the meaning of the independent words are displayed to the user. Displayed (presented). The user can select an independent word that can be a candidate for dictionary addition via the selection screen displayed by the display unit 43.

  Here, FIG. 17 shows an example of a selection screen displayed by the display unit 43 included in the user terminal device 40. As shown in FIG. 17, on the selection screen 431, the independent word (entry word) searched by the morphological analysis dictionary search unit 33, the part of speech of the independent word, and the meaning of the independent word are displayed. The selection screen 431 displays check boxes in association with the independent words searched by the morphological analysis dictionary search unit 33.

  By referring to the selection screen 431, the user can confirm the part of speech and the meaning of the independent word searched by the morphological analysis dictionary search unit 33. In addition, the user checks the check box displayed on the selection screen 431 by, for example, an operation on the input unit 44 included in the user terminal device 40, so that the independent words associated with the check box are dictionaryd. It can be selected as an independent word that can be a candidate for additional registration.

  In the dictionary evaluation support device 30 (the dictionary pattern creation unit 34 included in the dictionary evaluation support device 30), pattern matching is performed based on only the independent words selected by the user as described above among the independent words searched by the morphological analysis dictionary search unit 33. A pattern is created.

  As described above, in the present embodiment, an independent word having the same sound and the same part of speech as the independent word extracted by the regular expression analyzing unit 32 is searched by the morphological analysis dictionary searching unit 33, and the searched independent word, the independent word The part of speech of the word and the meaning of the independent word are output to the user terminal device 40. Thereby, in the present embodiment, the user can confirm the part of speech and meaning of the independent word on the selection screen on which the independent word, the part of speech of the independent word, and the meaning of the independent word are displayed. It is possible to easily determine an appropriate independent word.

  Further, in the present embodiment, by limiting the pattern matching pattern created by the dictionary pattern creation unit 34 to only independent words selected by the user, the processing efficiency in the dictionary evaluation support device 30 can be improved. It becomes possible.

[Fifth Embodiment]
Next, a fifth embodiment of the present invention will be described. The functional configuration of the dictionary evaluation support apparatus according to this embodiment is the same as that of the first embodiment described above, and will be described with reference to FIG. 2 described above.

  In the present embodiment, when the number of example sentences including the corresponding expression of the corresponding pattern and the exclusion pattern stored in the pattern match dictionary 24 is not less than a predetermined number, the number of example sentences is reduced (hereinafter, example sentence reduction). This is different from the first embodiment described above in that the processing and notation are executed.

  With reference to the flowchart of FIG. 18, the process procedure of the dictionary evaluation assistance apparatus 30 which concerns on this embodiment is demonstrated.

  First, the process of step S21-step S26 equivalent to the process of step S1-step S6 shown in FIG. 7 mentioned above is performed.

  The corresponding pattern included in the pattern information read from the pattern match dictionary 24 in step S21 is referred to as a target corresponding pattern. In addition, the exclusion pattern included in the pattern information read from the pattern match dictionary 24 in step S21 is referred to as a target exclusion pattern.

  In addition, the target corresponding pattern, the target excluded pattern, and the regular expression of the pattern matching pattern created in step S24 are analyzed, and each expression converted into a specific expression is referred to as a corresponding expression.

  Next, the result determination unit 36 executes the example sentence deletion process with reference to the pattern match result created by the pattern matching unit 35 (the pattern match result after the example sentence addition process in step S26) (step S27). In this example sentence deletion process, based on the morphological analysis result of the example sentence including the corresponding expression of the target corresponding pattern and the target excluded pattern, the example sentence (information regarding the target expression of the target corresponding pattern and the target excluded pattern) is converted into the pattern matching unit 35. Is deleted from the pattern match result created by. That is, in the example sentence deletion process, the number of example sentences (first example sentences) including the corresponding expressions of the target corresponding pattern and the target excluded pattern in the pattern matching result created by the pattern matching unit 35 is reduced. Details of the example sentence deletion process will be described later.

  When the process of step S27 is executed, the process of step S28 corresponding to the process of step S7 described above is executed. In this case, the result output unit 38 outputs the pattern match result from which the example sentence (information related to the example sentence is deleted) by the example sentence deletion process to the user terminal device 40 as dictionary evaluation support information.

  Next, a processing procedure of the above-described example sentence deletion process (the above-described process of step S27 shown in FIG. 18) will be described with reference to the flowchart of FIG.

  The example sentence deletion process is executed for each of the corresponding expressions of the target corresponding pattern and the target excluded pattern. Hereinafter, a corresponding expression that is a target of the example sentence deletion process is referred to as a target corresponding expression.

  First, the result determination unit 36 refers to the pattern matching result created by the pattern matching unit 35 and counts (number of) example sentences including the target corresponding expression. That is, the result determination unit 36 counts the number of example sentences (number of example sentences of the target corresponding expression) associated with the target corresponding expression in the pattern match result created by the pattern matching unit 35.

  Here, FIG. 20 shows an example of an example sentence including, for example, a corresponding expression “/ meeting [noun]” of the target corresponding pattern included in the pattern matching result created by the pattern matching unit 35.

  As shown in FIG. 20, the example sentences associated with the corresponding expression “/ meeting [noun]” of the target corresponding pattern in the pattern matching result are, for example, “meeting with customer on 8th”, “today's meeting”. Will be held in the meeting room from 15:00 "and" Scheduled to meet with customers in August. "

  Hereafter, the example sentence “Meeting with the customer on the 8th” shown in FIG. 20 will be referred to as Example sentence 1, and the example sentence “Today's meeting will be held in the meeting room from 15:00” will be referred to as Example sentence 2. "Schedule to meet with customer" inside is called example sentence 3.

  In the example illustrated in FIG. 20, the number of example sentences of the target corresponding expression “/ meeting [noun]” included in the pattern matching result is three.

  Returning to FIG. 19 again, the result determination unit 36 determines whether or not the number of example sentences of the target expression is within a predetermined number (second specified number) (below) (step S31). The second specified number indicates the upper limit of the target corresponding expression. Here, it is assumed that the second specified number (upper limit of the number of example sentences) is 2.

  For example, since the number of example sentences of the target corresponding expression “/ meeting [noun]” illustrated in FIG. 20 is 3, the result determination unit 36 determines that the number of example sentences of the target corresponding expression is not within the second specified number.

  When it is determined that the number of example sentences of the target corresponding expression is not within the second specified number (NO in step S31), the result determining unit 36 selects the target corresponding expression included in the pattern match result created by the pattern matching unit 35. Including example sentences (example sentences for the target corresponding expression) are sequentially extracted from the pattern match result. Here, it is assumed that the above-described example sentence 1, example sentence 2, and example sentence 3 are extracted in this order.

  Here, in the example sentence included in the pattern matching result generated by the pattern matching unit 35, the phrase including the corresponding expression and the constituent part of speech of the phrase immediately before and after the phrase are referred to as the corresponding constituent part of speech in the example sentence.

  The result determination unit 36 compares the corresponding component part of speech in the example sentence extracted from the pattern match result with the corresponding component part of speech in the already extracted example sentence to determine whether the corresponding component part of speech is the same or not. A determination is made based on the analysis result (step S32).

  As for the corresponding part of speech of the example sentence 1 “meeting with the customer on the 8th” shown in FIG. 20, the phrase immediately before the target corresponding expression (meeting) is “noun + particle”, and the phrase including the target corresponding expression is The “target corresponding expression” and the clause immediately after the target corresponding expression are “nouns”.

  In addition, the corresponding part of speech of Example sentence 2 “Today's meeting will be held in the meeting room from 15:00” shown in FIG. 20 is “Noun + particle” immediately before the target corresponding expression (meeting). The phrase including “is the target expression + particle”, and the immediately following phrase is “number”.

  Here, for example, assuming that the example sentence 1 has already been taken out, the corresponding component part of speech in the example sentence 2 is not the same as the corresponding constituent part of speech in the example sentence 1 already taken out. Therefore, the result determination unit 36 determines that the corresponding component part of speech in the example sentence 2 and the corresponding component part of speech in the already extracted example sentence 1 are not the same (NO in step S32).

  In this case, the result determination unit 36 determines whether or not the processing has been executed for all of the example sentences (here, example sentences 1 to 3) including the target corresponding expression (step S34).

  Here, since the process is not executed for the example sentence 3, the result determination unit 36 determines that the process is not executed for all the example sentences including the target corresponding expression (NO in step S34). In this case, it returns to above-mentioned step S32 and a process is repeated.

  Here, the corresponding component part of the example sentence 3 “scheduled to meet with customer in August” shown in FIG. 20 is that the phrase immediately before the target corresponding expression (meeting) is “noun + particle”, and the phrase including the corresponding expression is The phrase immediately after the “corresponding expression” and the target corresponding expression is “noun”.

  As described above, assuming that example sentence 1, example sentence 2, and example sentence 3 are extracted in this order, the corresponding component part of speech in example sentence 3 is the same as the corresponding component part of speech in example sentence 1 already extracted. Therefore, the result determination unit 36 determines that the corresponding component part of speech in the example sentence 3 and the corresponding component part of speech in the already extracted example sentence 1 are the same (YES in step S32).

  In this case, the result determination unit 32 deletes the example sentence 3 (information about the example sentence) from the pattern matching result created by the pattern matching unit 35 (step S33).

  Thereby, the result determination unit 32 can reduce the number of example sentences of the target corresponding expression in the pattern match result created by the pattern matching unit 35. When the process of step S33 is executed, the process of step S34 described above is executed.

  When it is determined in step S34 that the processing has been executed for all the example sentences including the target corresponding expression, the result determination unit 36 determines whether the example sentence deletion process has been executed for all the target expressions of the target corresponding pattern and the target exclusion pattern. Is determined (step S35).

  If it is determined that the example sentence deletion process has not been executed for all the corresponding expressions (NO in step S35), the process returns to step S31 and is repeated. In this case, the process is executed with the corresponding expression that has not been processed as the target corresponding expression.

  On the other hand, when it is determined that the example sentence deletion process has been executed for all the corresponding expressions (YES in step S35), the example sentence deletion process is terminated.

  If it is determined in step S31 that the number of example sentences is within the second specified number, the process of step S35 is executed. That is, since the number of example sentences of the target applicable expression is appropriate, the example sentence (information related to the target applicable expression) is not deleted.

  The example sentence deletion process shown in FIG. 19 has been described as being executed for each of the corresponding expressions of the target corresponding pattern and the target excluded pattern. However, for example, a configuration executed for the corresponding expression of the pattern matching pattern It does not matter.

  As described above, in the present embodiment, when the number of example sentences of the target applicable expression is not within the second specified number, the corresponding constituent part of speech is the same as the corresponding constituent part of speech in other example sentences (that is, other example sentences and meanings). Information on example sentences having similar structures is deleted from the pattern match result. Thereby, in this embodiment, when showing a pattern matching result to a user as dictionary evaluation assistance information, an appropriate number of example sentences can be shown.

  Note that, in the example sentence deletion process in the present embodiment, it has been described that the process is executed for all example sentences including the target applicable expression. For example, the number of example sentences of the target applicable expression is the second specified number (that is, the upper limit). If it is within the range, the example sentence deletion process may be terminated (stopped).

  Further, in the example sentence deletion process of the present embodiment, it has been described that it is determined whether or not to delete an example sentence by comparing the corresponding component parts of speech in the example sentence. For example, a specific expression of a phrase constituting the example sentence is used. The configuration may be such that the determination is performed. In addition, the range of the corresponding part of speech used for the determination (that is, the range of phrases to be compared) may be changed as appropriate.

[Sixth Embodiment]
Next, a sixth embodiment of the present invention will be described with reference to FIG. FIG. 21 is a block diagram mainly showing a functional configuration of the dictionary evaluation support apparatus 300 according to the present embodiment. The same parts as those in FIG. 2 described above are denoted by the same reference numerals, and detailed description thereof is omitted. Here, parts different from FIG. 2 will be mainly described.

  The hardware configuration of the dictionary evaluation support apparatus according to this embodiment is the same as that of the first embodiment described above, and will be described with reference to FIG. 1 as appropriate.

  This embodiment is different from the first embodiment described above in that an example sentence obtained by converting an example sentence stored in the example sentence corpus 22 is added in the example sentence addition process.

  As shown in FIG. 21, the dictionary evaluation support apparatus 300 includes an example sentence adding unit 39. In the present embodiment, the example sentence adding unit 39 is realized by the computer 10 illustrated in FIG. 1 executing the program 21 stored in the external storage device 20.

  The dictionary evaluation support device 300 includes a synonym dictionary 25. In the present embodiment, the synonym dictionary 25 is stored in, for example, the external storage device 20 (or a nonvolatile memory device).

  The synonym dictionary 25 stores, for example, synonyms of independent words stored in the morphological analysis dictionary 23 (words having similar meanings to the independent words). In the synonym dictionary 25, for example, “a little (adverb)” and “a little (adverb)” are stored as synonyms of “some (adverb)” which is an independent word.

  The example sentence adding unit 39 executes a process of adding an example sentence (third example sentence) including the expression represented by the target corresponding pattern to the example sentence corpus 22. The example sentence adding unit 39 includes an example sentence converting unit 391.

  The example sentence conversion unit 391 searches the synonym dictionary 25 for synonyms of the independent words (first independent words) extracted by the regular expression analysis unit 32.

  The example sentence conversion unit 391 acquires (searches) an example sentence (fourth example sentence) including the searched synonym from the example sentence corpus 22.

  The example sentence conversion unit 391 creates an example sentence by converting synonyms included in the acquired example sentence into independent words extracted by the regular expression analysis unit 32. That is, the example sentence conversion unit 391 creates an example sentence by converting the acquired example sentence.

  Further, the example sentence adding unit 391 displays a correct expression (hereinafter referred to as a correct expression) of the independent word specified by the user when the independent word extracted by the regular expression analyzing unit 32 is erroneously written, for example. get. In this case, the user can specify the correct expression by operating the input unit 44 included in the user terminal device 40.

  The example sentence adding unit 391 acquires (searches) an example sentence including the acquired correct expression from the example sentence corpus 22.

  The example sentence conversion unit 391 creates an example sentence by converting the correct expression included in the acquired example sentence into an independent word extracted by the regular expression analysis unit 32.

  The example sentence conversion unit 391 adds the created example sentence to the example sentence corpus 22. Further, the example sentence conversion unit 391 creates a pattern match result (pattern match result of the third example sentence) of the created example sentence (example sentence added to the example sentence corpus 22), and the pattern match created by the pattern matching part 35 Add to the result.

  Hereinafter, the operation of the dictionary evaluation support apparatus 300 according to the present embodiment will be described. Since the operation of the dictionary evaluation support apparatus 300 according to the present embodiment is the same as that of the first embodiment except for the example sentence addition process shown in FIG. 7 described above, detailed description thereof is omitted.

  With reference to the flowchart of FIG. 22, a description will be given of a processing procedure of example sentence addition processing in the dictionary evaluation support apparatus 300 according to this embodiment.

  The example sentence addition process is executed for each of the corresponding expressions of the target corresponding pattern and the target excluded pattern, as in the first embodiment. Hereinafter, the corresponding expression that is the target of the example sentence addition process is referred to as a target corresponding expression.

  In the following description, the independent word (first independent word) extracted by the regular expression analysis unit 32 is referred to as a target expression.

  First, the process of step S41 corresponding to the process of step S11 shown in FIG. 10 described above is executed. When it is determined that the number of example sentences of the target corresponding expression is greater than or equal to the first specified number (YES in step S41), the process in step S47 described later is executed.

  On the other hand, when it is determined that the number of example sentences of the target applicable expression is not equal to or greater than the first specified number (NO in step S41), the example sentence conversion unit 391 included in the example sentence adding unit 39 receives the target expression (regular expression analyzing unit 32). A synonym of the extracted self-supporting word) is searched from the synonym dictionary 25. Thereby, the example sentence conversion unit 391 determines whether or not the synonym of the target expression is in the synonym dictionary 25 (step S42).

  When it is determined that the synonym of the target expression is in the synonym dictionary 25 (YES in step S42), the example sentence conversion unit 391 acquires the synonym of the target expression from the synonym dictionary 25 (step S43). Specifically, for example, when the purpose expression is “some”, “a little” or the like is acquired as a synonym of the object expression.

  Next, the example sentence conversion unit 391 acquires (searches) an example sentence including the acquired synonym from the example sentence corpus 22 (step S44). Specifically, for example, when “a little” is acquired as a synonym of the target expression “some”, the example sentence “I am a little confident in sewing” is acquired.

  The example sentence conversion unit 391 converts a synonym included in the acquired example sentence into a target expression (step S45). Thereby, the example sentence conversion part 391 produces the example sentence by which the synonym was converted into the target expression.

  As described above, when the synonym of the target expression “some” is “little” and the example sentence acquired by the example sentence conversion unit 391 is “I am a little confident in sewing”, the example sentence conversion unit 391 The example sentence “I am somewhat confident in sewing” is created by converting the synonym “little” into the target expression “some”.

  The example sentence conversion unit 391 adds the created example sentence to the example sentence corpus 22, and adds information related to the example sentence to the pattern match result created by the pattern matching unit 35 (step S46). The information about the example sentence created by the example sentence conversion unit 391 includes the top-priority pattern (target corresponding pattern or target exclusion pattern) matched as described above, the corresponding expression (target corresponding expression), the example sentence, and the morphological analysis result of the example sentence. Is included. Note that the example sentence information generated by the example sentence conversion unit 391 (third pattern match result) is generated by the example sentence conversion unit 391 based on the target corresponding expression, the example sentence, and the like.

  Next, it is determined whether or not the example sentence addition process (the process after step S41 described above) has been executed for all the corresponding expressions of the target corresponding pattern and the target excluded pattern (step S47).

  When it is determined that the example sentence addition process has not been executed for all the corresponding expressions (NO in step S47), the process returns to step S41 and the process is repeated. In this case, the process is executed with the corresponding expression that has not been processed as the target corresponding expression.

  On the other hand, when it is determined that the example sentence addition process has been executed for all the corresponding expressions (YES in step S47), the example sentence addition process is terminated.

  In this case, the pattern match results to which the information about the example sentences created by the above-described example sentence conversion unit 391 is added are sorted by corresponding expressions (target corresponding patterns, target excluded patterns, and pattern matching pattern corresponding expressions). The pattern match result is presented to the user via the user terminal device 40 as dictionary evaluation support information.

  On the other hand, when it is determined in step S42 described above that the synonym of the target expression is not in the synonym dictionary 25, the screen for allowing the user to input (designate) the correct expression of the target expression as the target expression is erroneous. Are displayed on the user terminal device 40 (display unit 43 included therein). Thereby, the user can designate a correct expression by an operation on the input unit 44 included in the user terminal device 40, for example. Specifically, for example, when the purpose expression is “third party”, the user designates “third party” as the correct expression.

  The correct answer expression designated by the user is transmitted to the dictionary evaluation support device 300 via the communication interface 45 included in the user terminal device 40.

  The example sentence conversion unit 391 acquires the correct expression (correct expression specified by the user) transmitted by the user terminal device 40 (step S48).

  Next, the example sentence conversion unit 391 acquires (searches) an example sentence including the acquired correct expression from the example sentence corpus 22 (step S49). Specifically, for example, when “third party” is acquired as the correct expression of the target expression “third party”, the example sentence “Request examination from a third party organization” is acquired.

  The example sentence conversion unit 391 converts the correct expression included in the acquired example sentence into a target expression (step S50). As a result, the example sentence conversion unit 391 creates an example sentence in which the correct expression is converted into the target expression.

  As described above, when the correct expression of the target expression “third person” is “third party” and the example sentence acquired by the example sentence conversion unit 391 is “request examination from a third party organization”. The example sentence conversion unit 391 creates the example sentence “Request examination from a third party organization” by converting the correct expression “third party” into the target expression “third party”. When the process of step S50 is executed, the process of step S46 described above is executed.

  As described above, in the present embodiment, a synonym of the target expression (an independent word extracted by the regular expression analysis unit 32) is searched from the synonym dictionary 25, and an example sentence including the searched synonym is acquired from the example sentence corpus 22. The In this embodiment, an example sentence is created by converting a synonym included in an example sentence acquired from the example sentence corpus 22 into a target expression, the created example sentence is added to the example sentence corpus 22, and the created example sentence Information is added to the pattern match result.

  On the other hand, in this embodiment, when the synonym of the target expression is not in the synonym dictionary 25, the user specifies (inputs) the correct expression of the target expression, and an example sentence including the correct expression is acquired from the example sentence corpus 22. The In the present embodiment, an example sentence is created by converting a correct expression included in an example sentence acquired from the example sentence corpus 22 into a target expression, the created example sentence is added to the example sentence corpus 22, and the created example sentence Information about is added to the pattern match result.

  Thus, in the present embodiment, an example sentence can be created by converting the synonym or the correct expression into a target expression in the example sentence including the correct expression for the synonym or word notation, and thus the first embodiment described above. As described in the above, for example, it is possible to add an example sentence correct as a syntax at a low cost compared to a case where an example sentence is acquired from the Web space.

  Moreover, in this embodiment, since an example sentence can be added easily as mentioned above, the creation cost of the dictionary creation assistance information for assisting a user's dictionary evaluation work can be reduced, for example.

  Note that the present invention is not limited to the above-described embodiments as they are, and can be embodied by modifying the constituent elements without departing from the scope of the invention in the implementation stage. Further, various inventions can be formed by appropriately combining a plurality of constituent elements disclosed in the above embodiments. For example, some components may be deleted from all the components shown in each embodiment. Furthermore, constituent elements over different embodiments may be appropriately combined.

  DESCRIPTION OF SYMBOLS 10 ... Computer, 20 ... External storage device, 22 ... Example sentence corpus, 23 ... Morphological analysis dictionary, 24 ... Pattern match dictionary, 25 ... Synonym dictionary, 30, 300 ... Dictionary evaluation support device, 31 ... Morphological analysis part, 32 ... Regular Expression analysis unit, 33 ... Morphological analysis dictionary search unit, 334 ... Dictionary pattern creation unit, 35 ... Pattern matching unit, 36 ... Result determination unit, 37, 39 ... Example sentence addition unit, 38 ... Result output unit, 371 ... Example sentence search unit 372: Example sentence evaluation unit, 391: Example sentence conversion unit.

Claims (8)

An example sentence corpus that stores a plurality of example sentences including independent words in advance;
A morphological analysis dictionary that stores in advance independent words, readings of the independent words, and parts of speech of the independent words;
A pattern match dictionary that pre-stores a first pattern representing an expression including a first independent word;
Extracting means for extracting a first independent word included in an expression represented by the first pattern by analyzing the first pattern stored in the pattern match dictionary;
A morpheme analysis dictionary search unit that searches the morpheme analysis dictionary for a second independent word that has the same reading and part of speech as the extracted first independent word;
Pattern creating means for creating a second pattern representing an expression including the second independent word based on the first pattern stored in the pattern match dictionary and the searched second independent word;
By applying the first pattern stored in the pattern match dictionary to the example sentence stored in the example sentence corpus, the pattern match result of the first example sentence including the expression represented by the first pattern is obtained. First pattern matching means to create;
By applying the second pattern created by the pattern creating means to the example sentence stored in the example sentence corpus, the pattern matching result of the second example sentence including the expression represented by the second pattern is created. Second pattern matching means to:
Determining means for determining whether the number of first example sentences including the expression represented by the first pattern stored in the pattern match dictionary is equal to or greater than a predetermined number;
If it is determined that the number of the first example sentences is not equal to or greater than a predetermined number, a third example sentence including an expression represented by the first pattern stored in the pattern match dictionary is obtained from an external corpus. An adding means for searching and adding the third example sentence to the example sentence corpus;
The pattern matching result of the first example sentence created by the first pattern matching means, the pattern matching result of the second example sentence created by the second pattern matching means, and the expression represented by the first pattern And a result output means for outputting a pattern match result of the third example sentence created on the basis of the third example sentence. The additional means includes
Example sentence search means for searching a third example sentence including an expression represented by the first pattern stored in the pattern match dictionary from an external corpus;
Syntactic analysis of the third example sentence searched by the example sentence search means, and example sentence evaluation means for determining whether the third example sentence matches a predetermined condition based on the syntax analysis result,
2. The dictionary according to claim 1, wherein when it is determined that the third example sentence searched by the example sentence search means matches a predetermined condition, the third example sentence is added to the example sentence corpus. Evaluation support device. Further comprising a synonym dictionary for storing synonyms of independent words stored in the morphological analysis dictionary;
The additional means includes
Synonym dictionary searching means for searching the synonym of the extracted first independent word from the synonym dictionary;
Obtaining means for obtaining, from the example sentence corpus, a fourth example sentence including the synonym searched by the synonym dictionary searching means;
And example sentence conversion means for creating a third example sentence by converting a synonym included in the fourth example sentence acquired by the acquisition means into the extracted first independent word;
The dictionary evaluation support apparatus according to claim 1, wherein the third example sentence created by the example sentence conversion unit is added to the example sentence corpus. A further reduction means,
The example corpus stores morphological analysis results of the plurality of example sentences;
The determination means determines whether the number of first example sentences including an expression represented by the first pattern stored in the pattern match dictionary is equal to or less than a predetermined number;
Said reducing means, when the number of the first example sentence is determined not less than the number of predetermined said respective first, based on the morphological analysis result of the respective first example sentence stored in the example sentence corpus By comparing the phrase including the expression represented by the first pattern in one example sentence and the constituent parts of speech of the phrase before and after the phrase , the first in the pattern matching result created by the first pattern matching means dictionary evaluation support apparatus of claim 1, wherein removing the first example sentence. The morpheme analysis dictionary further stores a standard expression of independent words stored in the morpheme analysis dictionary,
2. The dictionary evaluation support according to claim 1, wherein the morpheme analysis dictionary search unit searches the morpheme analysis dictionary for a second independent word having the same standard notation as the extracted first independent word. apparatus. The morphological analysis dictionary further stores the meaning classification of the independent words stored in the morphological analysis dictionary,
The dictionary evaluation support according to claim 1, wherein the morpheme analysis dictionary search means searches the morpheme analysis dictionary for a second independent word having the same semantic classification as the extracted first independent word. apparatus. The morpheme analysis dictionary further stores semantic information indicating the meaning of independent words stored in the morpheme analysis dictionary,
The morpheme analysis dictionary search means outputs semantic information indicating the meaning of the second independent word searched by the morpheme analysis dictionary search means and the second independent word stored in the morpheme analysis dictionary,
The corresponding pattern creating means creates a second pattern representing an expression including the second independent word designated by the user among the second independent words output by the morphological analysis dictionary searching means. The dictionary evaluation support device according to claim 1. An example sentence corpus that stores a plurality of example sentences including independent words in advance, a morphological analysis dictionary that stores in advance independent words, readings of the independent words, and parts of speech of the independent words, and a first representing an expression including the first independent word In a dictionary evaluation support device composed of an external storage device having a pattern match dictionary that prestores the pattern and a computer using the external storage device, a program executed by the computer,
In the computer,
Analyzing a first pattern stored in the pattern match dictionary to extract a first independent word included in an expression represented by the first pattern;
Searching the morphological analysis dictionary for a second independent word whose reading and part of speech are the same as the extracted first independent word;
Creating a second pattern representing an expression including the second independent word based on the first pattern stored in the pattern match dictionary and the searched second independent word;
By applying the first pattern stored in the pattern match dictionary to the example sentence stored in the example sentence corpus, the pattern match result of the first example sentence including the expression represented by the first pattern is obtained. A step to create,
Applying the generated second pattern to an example sentence stored in the example sentence corpus to create a pattern match result of a second example sentence including an expression represented by the second pattern;
Determining whether the number of first example sentences including an expression represented by the first pattern stored in the pattern match dictionary is greater than or equal to a predetermined number;
If it is determined that the number of the first example sentences is not equal to or greater than a predetermined number, a third example sentence including an expression represented by the first pattern stored in the pattern match dictionary is obtained from an external corpus. Searching and adding the third example sentence to the example sentence corpus;
The pattern matching result of the created first example sentence, the pattern matching result of the created second example sentence, the expression represented by the first pattern, and the third example sentence created based on the third example sentence And a step for outputting the pattern matching result of the example sentence of 3.

Images