JP2011118861A - Device, program and method for checking document - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a document check device which allows checking the consistency of using a specific character string in a document. <P>SOLUTION: The document check device 100 includes a memory 124 for storing document data. The memory 124 stores, when a term following the specific character string is used to show that it has already appeared in the document, information of the specific character string and information of a plurality of term suffixes to be used in common at the end of the term. A document data analysis program 131 recognizes and specifies a specific character string and term suffixes in the document, performs a morphological analysis of the document, specifies parts of speech in the document, and connects a noun that continues before term suffixes to the term suffixes on the basis of a result of the morphological analysis to thereby specify term candidates being candidates for the term. The consistency of using a specific character string is checked among the specific term documents. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a document check apparatus, a document check program, and a document check method for checking the consistency of descriptions in a document.

  Some documents require particularly stringent consistency in the description. For example, official documents, patent documents, especially descriptions of the scope of claims, and Japanese translations of English contracts.

  In such a document, the accuracy of the description, correctness of the description format, consistency of description, etc. are particularly important. For example, in patent specifications, the accuracy of the description is checked. A technique relating to software has been proposed (see, for example, Patent Document 1).

  However, in this patent document 1, the correctness of the description format, the check by the syntax analysis of the sentence, the check of the consistency of the used codes, etc. are merely performed.

  In addition, a computer is used to search all specified codes from documents such as patent application specifications, and for each code, data indicating the position and a certain length adjacent to the code. A technique is also disclosed in which a character string is automatically extracted and the extraction results are sorted and displayed in code order (see, for example, Patent Document 2).

  In the invention disclosed in Patent Document 2, the extracted character string includes a word with a sign, which is displayed side by side, so that it is possible to easily inspect a signing error and a difference in expression of the word.

  Alternatively, Patent Document 3 discloses a technique for efficiently inputting an element name and a code that follows the element name so that there is no mismatch. In the technique disclosed in Patent Document 3, the following processing is performed. When an appropriate operation is performed while editing a document such as a word processor, (1) an element name (for example, “wheel”) is extracted from the character string input at the editing position of the document. (2) The extracted element name is searched from the document. (3) A character string including the searched position is extracted from the document and analyzed, and a code (for example, “6”, “7”) corresponding to the element name extracted in (1) is acquired. (4) When a plurality of codes are obtained, the codes are displayed side by side on the screen so that they can be selected. (5) The code selected by the user is automatically input at the editing position of the document.

  However, these techniques are merely techniques for detecting a mismatch between a code of a patent application specification and a character string corresponding thereto, or a technique for supporting an input operation to match.

JP 2002-183278 A JP 09-259148 A JP 2005-25265 A

  However, in the types of documents described above, unique terms may be used in order to ensure the strictness of the document as compared with the description in a normal sentence. For example, if a term that appears in the description of a document refers to the same term that has already been described in the document, when the term appears in the document for the second time or later, There is a habit of demonstrating that the term has already appeared by prepending the word.

  For example, in the claims, strictness is particularly required for the use of “above”. In the description of the scope of claims, “this” or the like may be used to perform the same function as “the above”. In this case, in general, the term “subject” is often used with the intention of referring to the term that has just appeared, not just to indicate that it has already occurred.

  Whether this “above” (“said” in English) precedes the term is not just a formal issue, but can directly affect the interpretation of the scope of rights. For example, in the US case of Bell Communications Research Inc. v. Vitalink Communications Corp. (55 F.3d 615, 34 US.SPQ2d (BNA) 1816 (Fed. Cir.1995)) The description of “packet” described in the premise part (preamble part) of the claim by the description of “the packet” limits the technical content of “packet” in the main part of the claim. There is an example of interpretation of.

  In addition, a name that is very similar may be used by mistake when calling the same target component, and both may be mixed in the claims. For example, in the case where the description “○ △ □ ☆ ○ □ means” has already been mentioned, the description “the above-mentioned ○ Δ □□ ○ □ means” may exist after that. In this case, for the drafter of the claims, such a description is made with the recognition that “means of △△ □□ ○ □” has already been issued. However, to be precise, “○ △ □□ ○ □ Means” has not been issued, so in the examination, the description of the claim is unclear (violation of Patent Act Article 36 (6) (ii)), After the patent was granted, there was a problem that it could affect the interpretation of the scope of rights.

  The present invention has been made in order to solve the above-described problems. The object of the present invention is to identify a specific character string in a document and a term following the character string in the document. It is to provide a document check apparatus, a document check program, and a document check method capable of checking the consistency of the use of such a specific character string when used to indicate that it is.

  According to one aspect of the present invention, there is provided a document checking device for checking the consistency of descriptions in a document to be analyzed. The document check device includes a storage unit that stores document data representing a document, and a specific character string is used to indicate that a term subsequent to the specific character string has already appeared in the document. Information acquisition means for acquiring information on a specific character string and part of speech specification means for specifying a part of speech included in the document are provided. The part-of-speech identifying means is a means for recognizing and identifying a specific character string based on the information acquired by the information acquiring means in the document, and performing morphological analysis on the document, and the part-of-speech in the document Morphological analysis means for specifying Based on the result of the morphological analysis, the document check device is connected between the term candidate recognition means for identifying the term candidates that are term candidates by connecting consecutive nouns in the document and the identified term candidates. There is further provided consistency checking means for checking the consistency of use of the character string and display control means for causing the display device to display the consistency check result.

  Preferably, the information acquisition means further acquires information on a plurality of term suffixes commonly used for the term ending. Based on the result of morphological analysis, the term candidate recognizing means identifies a term candidate that is a term candidate by linking a noun that precedes the term suffix in the document to the term suffix.

  Preferably, the information acquisition unit further includes a user dictionary in which a term candidate that has not been specified at the time of the process of specifying a term candidate in the term candidate recognition unit in the document is registered as a term candidate by user selection. Get information. The term candidate recognizing means further identifies the term candidate with reference to the user dictionary.

  Preferably, the storage unit stores information on a specific character string, information on a plurality of term suffixes, and a user dictionary. The information acquisition unit reads information on a specific character string, information on a plurality of term suffixes, and information on a user dictionary from the storage unit.

  Preferably, the documents are classified into a plurality of groups for each content field. The user dictionary is divided into partial dictionaries for each group, and the user registers unspecified term candidates in the partial dictionary corresponding to the content field of the document.

  Preferably, the storage means further stores in advance a specific prefix to be prefixed to the term. The term candidate recognition unit further concatenates a specific prefix if a specific prefix continues before the term after the concatenation of a noun that precedes the term suffix to the term suffix. Thus, the term candidate is specified.

  Preferably, information on a specific character string, information on a plurality of term suffixes, and a user dictionary are stored in an external storage device outside the document check device. The information acquisition means acquires information on a specific character string, information on a plurality of term suffixes, and information on a user dictionary from an external storage device by communication.

  Preferably, the document data includes first partial document data representing a first document used to indicate that a specific character string is a term subsequent to the specific character string. And a second partial document for explaining a content defined by the first partial document data and expressing a second document used by adding a description term corresponding to the term to a reference numeral Data. The morpheme analysis unit performs morpheme analysis in common with the first and second partial document data. The document check apparatus further includes a code check unit for checking the consistency between the explanation term and the code for the second partial document data.

  According to another aspect of the present invention, there is provided a document check program for causing a computer including an arithmetic device and a storage device that stores document data representing a document to check consistency of descriptions in a document to be analyzed. Provided. A document check program obtains information about a specific character string when the arithmetic unit is used to indicate that the specific character string is a term that follows the specific character string. And a step of specifying a part of speech included in the document. The step of specifying the part of speech includes a step in which the arithmetic device recognizes and specifies a specific character string based on the acquired information in the document, and the arithmetic device performs morphological analysis on the document, Identifying the part of speech within. The document check program includes a step in which a computing device connects consecutive nouns in a document based on the result of morphological analysis to identify term candidates that are term candidates, and the computing device uses the specified term The method further includes a step of checking the consistency of use of the specific character string between the candidates, and a step of causing the display device to display a result of the consistency check on the arithmetic device.

  Preferably, the step of acquiring information includes the step of acquiring information of a plurality of term suffixes commonly used for the ending of the term. The step of identifying term candidates includes the step of identifying term candidates that are term candidates by connecting consecutive nouns to the term suffix in the document based on the result of morphological analysis. Including.

  Preferably, the obtaining step further includes a user who has registered a term candidate that has not been identified as a term candidate by user selection at the time of the process of identifying the term candidate in the step of identifying the term candidate in the document. Including obtaining dictionary information. The step of specifying the term candidates further includes the step of specifying the term candidates with reference to the user dictionary.

  Preferably, the storage device stores information on a specific character string, information on a plurality of term suffixes, and a user dictionary. The obtaining step includes a step of reading out information on a specific character string, information on a plurality of term suffixes, and information on a user dictionary from the storage device.

  Preferably, the documents are classified into a plurality of groups for each content field. The user dictionary is divided into partial dictionaries for each group, and the user registers unspecified term candidates in the partial dictionary corresponding to the content field of the document.

  Preferably, the storage device further stores in advance a specific prefix that is prefixed to the term. The step of identifying a term candidate includes adding a specific prefix if a specific prefix continues before the term after the concatenation of a noun that precedes the term suffix to the term suffix. It includes the step of identifying term candidates by concatenation.

  Preferably, information on a specific character string, information on a plurality of term suffixes, and a user dictionary are stored in an external storage device outside the computer on which the document check program is executed. The step of acquiring includes a step of acquiring information on a specific character string, information on a plurality of term suffixes, and information on a user dictionary from an external storage device by communication.

  Preferably, the document data includes first partial document data representing a first document used to indicate that a specific character string is a term subsequent to the specific character string. And a second partial document for explaining a content defined by the first partial document data and expressing a second document used by adding a description term corresponding to the term to a reference numeral Data. The step of specifying the part of speech includes the step of performing morphological analysis in common with the first and second partial document data. The document check process further includes a step of checking the consistency between the explanation term and the code for the second partial document data.

  According to another aspect of the present invention, there is provided a document check method for causing a computer including an arithmetic device and a storage device that stores document data representing a document to check consistency of descriptions in a document to be analyzed. Provided. The document check method is a method in which an arithmetic unit obtains information on a specific character string when a specific character string is used to indicate that a term following the specific character string has already appeared in the document. And a step of specifying a part of speech included in the document. The step of specifying the part of speech includes a step in which the arithmetic device recognizes and specifies a specific character string based on the acquired information in the document, and the arithmetic device performs morphological analysis on the document, Identifying the part of speech within. The document check method includes a step in which a computing device connects consecutive nouns in a document based on a result of morphological analysis to identify a term candidate that is a candidate for a term, and the computing device uses a specified term. The method further includes a step of checking the consistency of use of the specific character string between the candidates, and a step of causing the display device to display a result of the consistency check on the arithmetic device.

  Preferably, the step of acquiring information includes the step of acquiring information of a plurality of term suffixes commonly used for the ending of the term. The step of identifying term candidates includes the step of identifying term candidates that are term candidates by connecting consecutive nouns to the term suffix in the document based on the result of morphological analysis. Including.

  Preferably, the obtaining step further includes a user who has registered a term candidate that has not been identified as a term candidate by user selection at the time of the process of identifying the term candidate in the step of identifying the term candidate in the document. Including obtaining dictionary information. The step of specifying the term candidates further includes the step of specifying the term candidates with reference to the user dictionary.

  Preferably, the document data includes first partial document data representing a first document used to indicate that a specific character string is a term subsequent to the specific character string. And a second partial document for explaining a content defined by the first partial document data and expressing a second document used by adding a description term corresponding to the term to a reference numeral Data. The step of specifying the part of speech includes the step of performing morphological analysis in common with the first and second partial document data. The document check method further includes a step of checking the consistency between the explanation term and the sign for the second partial document data.

  In a document to be parsed, it is easy to confirm whether a particular character string is appropriate for the mode of use used to indicate that the term following the particular character string has already appeared in the document. It becomes possible to check.

  Moreover, even if the term is not registered in the user dictionary one by one, the term is specified using information on a plurality of term suffixes commonly used at the end of the term. It is possible to check the document with a significant reduction.

  In addition, in a document in which “specific term”, “specific character string” indicating that the specific term has already appeared, “descriptive term”, and “sign” attached to the explanatory term are used, It becomes possible to check the consistency of use.

  Alternatively, when a candidate for a specific term is extracted, using a “term suffix” as a clue can be selected by a user setting.

It is a figure showing the structure of the document check apparatus 100 which concerns on Embodiment 1 in a block diagram format. It is a figure which shows the functional structure of the document check apparatus 100 which concerns on Embodiment 1 in a block diagram format. 6 is a flowchart for explaining the operation of the document check apparatus 100 according to the first embodiment. FIG. 4 is a flowchart for explaining in more detail the processing in step S106 and step S108 in FIG. 3. FIG. It is a figure which shows the list of special parts of speech. It is a figure which shows the symbol allocated by the component recognition process. It is a 1st figure for demonstrating the process in each process step of FIG. It is a 2nd figure for demonstrating the process in each process step of FIG. It is a figure which shows the example of the dictionary which registered the term equivalent to the part of speech P0 among the component element suffixes which are user dictionaries. It is a figure which shows the example of the dictionary which registered the term corresponding to the part of speech P1 among the component element suffixes which are a common dictionary. Among the common dictionaries, a dictionary in which terms corresponding to the part of speech P2 and the part of speech P3 are registered, a dictionary in which terms corresponding to the component prefixes part of speech H1, H2, H3 are registered, and a suffix TT. It is a figure which shows the example of the dictionary which registered the term. It is a figure which shows the example of the dictionary which each registered the term corresponding to suffix FT, compulsory noun FN, check character string ZZ, non-noun XN, and non-prefix XS among common dictionaries. It is a figure which illustrates notionally the correction process and component recognition process which a document data analysis part 120.2 performs. It is a figure which shows the example of description of the claim displayed by step S110 of FIG. It is a figure which shows the example of the whole screen displayed by step S110 of FIG. FIG. 10 is a flowchart for explaining the operation of the document check apparatus according to the second embodiment, and is a diagram for comparison with FIG. 3 according to the first embodiment. It is a flowchart for demonstrating the process of step 206 of FIG. It is a flowchart for demonstrating in detail the flow of step S208 and S210 of FIG. It is a figure which shows the structure of the dictionary of Embodiment 1 and Embodiment 2 in contrast. It is a figure explaining the "assignment symbol" assigned to the character string in the document of a check object. It is a figure explaining the allocation symbol allocated by front end processing. It is a figure explaining the allocation symbol in a morphological analysis. 4 is a table for explaining backend processing 1; 10 is a table for explaining correction processing 1; It is a table explaining the correction process 2. FIG. It is a table for demonstrating "the recognition process of a component." It is a table explaining the back end process 2. FIG. 10 is a table for explaining backend processing 3; In step 212, it is a figure which shows the example of a display which the data display analysis part 120.3 performs.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the following description, the same parts are denoted by the same reference numerals. Their names and functions are also the same. Therefore, detailed description thereof will not be repeated.

(Overview)
When document check apparatus 100 according to the embodiment is used in document data given from the outside, a specific character string is used to indicate that a term following the character string has already appeared in the document. In addition, a word / phrase which is a candidate for a term is extracted from the document, and it is checked whether or not the use of such a specific character string is consistent among the extracted candidates.

Here, it is necessary to pay attention to the following points.
1) The term preceded by a specific character string in the document is generally a noun or a noun phrase. However, for all nouns and noun phrases used in a document, a specific character string is not necessarily prefixed when used in the document for the second time or later. For example, nouns and noun phrases used to represent general concepts are not prefixed with a specific character string even after the second time. In other words, a specific character string is used on the assumption that the following term refers to a specific thing.

  2) On the other hand, in a document in a specific field (hereinafter referred to as “document field”), the ending of a term (hereinafter referred to as “specific term”) preceded by a specific character string as described above. There is a high probability of becoming a noun of a certain character string (hereinafter referred to as “term suffix”). For example, “period”, “document”, and the like.

  Further, even in the same document field, there are differences in the frequency of use of nouns and noun phrases representing constituent elements depending on the field of contents to be described in the document (hereinafter, “content field”).

  For example, taking the scope of claims as an example of the document field, in the description, “above”, “related” or the like is used as the specific character string. In the description of the claims, the specific term as described above is considered to represent the specific thing itself or an element constituting the specific thing. It will be referred to as a “component” as representing a specific type. At this time, the suffix of the component to which “the above” or the like is attached (hereinafter referred to as “component suffix” indicating a specific type of term suffixes) includes “means”, “apparatus”, “ There are many cases in the case of specific character strings such as “element” and “signal”. If such a term suffix (or component suffix) is registered in advance as a dedicated dictionary, it becomes a clue when a candidate for a specific term (or component) is extracted from a document. Regardless of the content field, a dedicated dictionary for term suffixes that are used more than a certain frequency will be referred to as a “common dictionary”.

  In the description of the scope of claims, nouns and noun phrases used as constituent elements tend to be different depending on the technical field covered by the scope of claims or the business scope of the applicant applying for the patent.

  3) As mentioned above, the term suffix (more specifically, the component suffix) is often common among terms, but not all “specific terms should be preceded by a specific character string”. ”Does not always have such a common ending. Furthermore, especially in the description of claims, new terms tend to be used from time to time as technology advances, so they can be registered in advance as “component suffixes”. Terms are not necessarily used within the scope. Therefore, when a candidate for a specific term (or component) is extracted using a common dictionary, there is a dictionary in which a user can register a candidate for a specific term (or component) that is missing from the extraction at any time. It is desirable to exist. Such a dictionary is called a “user dictionary”.

  In addition, since the frequency of the term used for every content field differs, the document check apparatus 100 can also be provided with a some dictionary corresponding to every group of the document by such a content field as a user dictionary. . In particular, in the description of the scope of claims, as described above, nouns and noun phrases used as components tend to be different in the business scope of the applicant. Alternatively, a different “customer dictionary” may be provided for each applicant. Of course, the document checking apparatus 100 may be configured to include a “technical field dictionary” for each technical field, not for each applicant.

  4) As described above, when candidate terms (or candidate constituent elements) are extracted, i) for each term candidate, i) for a term prefixed with a specific character string, Ii) For those that have not been prefixed with a specific character string, check whether the term candidate is first appearing in the document. It will be possible to check whether they are consistent.

  Here, in particular, in the description of the scope of claims, whether it is already appearing or first appearing should be determined only within the scope of the description of the claim if it is an independent claim. However, in the dependent claims, not only the description in each claim but also the description of the preceding claims on which the claims depend is followed in order until the description of the independent claims that will ultimately become the basis of the dependency. It is necessary to judge whether it has already appeared or first appearance within the range.

  Due to the circumstances as described above, in order to recognize the constituent elements from the sentence of the claims, information on the “word range and its part of speech” of the character string in the sentence is necessary. For this purpose, software of “Morphological Analysis Engine” can be used.

  Here, “morphological analysis” is natural language processing using a computer such as a computer, and includes grammatical information (gathering of grammar rules) of a target language and a corpus dictionary (a word list with information such as parts of speech). Is a process of dividing a sentence written in a natural language into columns of morphemes (Morpheme, the smallest unit meaningful in the language) and discriminating each part of speech.

  However, in the sentence of the claims, the morphological analysis engine generally only divides the sentence into morphemes and performs the process of specifying the part of speech of each morpheme, so that it is not sufficient for specifying the constituent elements. Therefore, it is necessary to use a dedicated dictionary as described above that stores partial words of the constituent elements.

  Japanese “Morphological Analysis Engine (MAE)” is already available as free software, for example, the following.

i) KAKASI “kanji kana simple inverter”, http://kakasi.namazu.org/
ii) MeCab (Mekabu), http://mecab.sourceforge.net/
iii) ChaSen, http://chasen-legacy.sourceforge.jp/
[Embodiment 1]
(Hardware configuration)
A hardware configuration of the document checking apparatus 100 according to the first embodiment will be described with reference to FIG. FIG. 1 is a block diagram showing the configuration of the document check apparatus 100 according to the first embodiment.

  Hereinafter, as an example of the document check, the document check apparatus 100 will be described assuming that the scope of the claims is a check target.

  The document checking apparatus 100 includes a computer main body 102, a monitor 104 as an output device or a display device, a keyboard 110 as an input device, and a mouse 112 as an input auxiliary device. The monitor 104, keyboard 110, and mouse 112 are connected to the computer main body 102 via the bus 105.

  The computer main body 102 is a flexible disk (Flexible Disc, hereinafter referred to as “FD”) drive 106 that is a reading device for an external recording medium, an optical disk drive 108 that is a reading device for another external recording medium, and an arithmetic processing device. A CPU (Central Processing Unit) 120, a memory 122 as a storage device, a direct access memory device as a mass storage device, for example, a hard disk 124, and a communication interface 128 as a communication device are included. These components are connected to each other by a bus 105.

  The FD drive 106 reads and writes information from and to the FD 116. The optical disc drive 108 reads information on an optical disc such as a CD-ROM (Compact Disc Read-Only Memory) 118. The communication interface 128 exchanges data with the outside.

  The CD-ROM 118 may be another medium, such as a DVD-ROM (Digital Versatile Disc) or a memory card, as long as it can record information such as a program installed in the computer main body. In this case, the computer main body 102 is provided with a drive device that can read these media.

  The memory 122 includes a ROM (Read Only Memory) and a RAM (Random Access Memory).

  The hard disk 124 includes a display control program 131, a document data analysis program 132, group information 133 representing a group of content fields of a document to be analyzed, common dictionary data 134, customer dictionary data 135, and a document to be checked. Data 136 is stored.

  The group information 133 is not particularly limited. For example, the group information 133 can be configured to be input by the user when starting the software.

  The common dictionary data 134 and the customer dictionary data 135 are stored in a storage device of another computer connected via a network, and the document check device 100 accesses these data via the communication interface 128. Thus, it is possible to adopt a configuration in which reading or writing is performed.

  The display control program 131 controls display of a screen serving as an interface between the document check apparatus 100 and the user. Therefore, this is for the process of controlling the process of displaying a screen for prompting input from the user or displaying the check result.

  As will be described later, the document data analysis program 132 extracts nouns and noun phrases as component candidates from the document data 136 to be checked based on the common dictionary data 134 and the customer dictionary data 135. Further, the document data analysis program 132 is for processing for checking the consistency of the use of a specific character string such as “above” or “related” for the component candidate extracted in this way. .

  Here, in the process in which the document data analysis program 132 extracts candidate components, the morphological analysis engine described above is used. Therefore, although not shown, the hard disk 124 also stores “grammar information of the target language” and “corpus dictionary” required when the morphological analysis engine performs the morphological analysis processing.

  The display control program 131 and the document data analysis program 132 may be supplied by being recorded on a storage medium such as the FD 116 or the CD-ROM 118, or may be supplied via a communication interface 128 by another computer. May be.

  As described above, the common dictionary data 134 is a dedicated dictionary for component suffixes that are used at a certain frequency regardless of the content field. In the example shown in FIG. 1, the content field exemplifies a case where the content field varies depending on the business scope of the applicant (customer). In this case, the component that is used at a certain frequency or more regardless of the customer. This is a dedicated dictionary for suffixes.

  On the other hand, the customer dictionary data 135 is a user dictionary that allows a user to register a candidate for a component missing from the extraction at any time when a noun or noun phrase that is a candidate for the component is extracted using a common dictionary. It is.

  Document data 136 is data of a document to be checked. Although not particularly limited, for example, when the document to be checked is a file created by specific word processor software, the document data 136 can be data obtained by extracting text data from this file. The display control program 131 controls the display of the document data on the document check apparatus 100 so that the display is performed in a predetermined layout from the information included in the text data. With such a configuration, the document checking apparatus 100 directly operates a file created by the word processor software and does not write data or the like, so that unexpected changes or the like are present in the file data. In addition, it is possible to prevent so-called “garbled characters” and “unexpected changes in layout information”.

  The CPU 120 functioning as an arithmetic processing unit executes processing corresponding to each program described above using the memory 122 as a working memory.

  The display control program 131 and the document data analysis program 132 are software executed by the CPU 120 as described above. Generally, such software is stored and distributed in a storage medium such as the CD-ROM 118 and the FD 116, read from the storage medium by the optical disk drive 108 or the FD drive 106, and temporarily stored in the hard disk 124. Alternatively, when the document check apparatus 100 is connected to a network, the document check apparatus 100 is temporarily copied to the hard disk 124 from a server on the network. Then, the data is further read from the hard disk 124 to the RAM in the memory 122 and executed by the CPU 120. In the case of network connection, the program may be directly loaded into the RAM and executed without being stored in the hard disk 124.

  The computer hardware itself shown in FIG. 1 and its operating principle are general. Therefore, an essential part for realizing the functions of the present invention is software stored in a storage medium such as the FD 116, the CD-ROM 118, and the hard disk 124.

(Functional configuration)
FIG. 2 is a block diagram showing the functional configuration of the document check apparatus 100 according to the first embodiment.

  FIG. 3 is a flowchart for explaining the operation of the document check apparatus 100 according to the first embodiment.

  With reference to FIG. 2 and FIG. 3, the functional configuration and operation of the document check apparatus 100 according to the first embodiment will be described.

  The document check apparatus 100 includes a hard disk 124 as a storage device, a monitor 104 as a display device, a keyboard 110 as an input device, and functional blocks of processing executed by the CPU 120. I-1) a document fetch unit 120.1 that acquires document data from a storage device such as the hard disk 124, and i-2) a document data analysis unit 120.2 that analyzes and checks document data, and ii) A data display output control unit 120.3 that displays on the monitor 104 a check result of the document data analysis unit 120.2 and an interface screen between the document data analysis unit 120.2 and the user, which are executed by the display control program 131. Prepare.

  In FIG. 2, customer dictionary data 135, which is a user dictionary, includes partial dictionaries 135.1 to 135.135 for each customer (for each business scope of the customer) corresponding to the content field of the document data. It is explicitly shown that it is divided into n.

  That is, the dedicated dictionary is classified into two types of dictionaries (customer dictionary and common dictionary). For example, a P0 dictionary described later is assigned to the customer dictionary, and a dictionary other than P0 is assigned to the common dictionary. For example, a plurality of users are divided into one large group composed of a plurality of small groups, and a customer dictionary is arranged for each small group and a common dictionary is arranged in the large group.

  With such a configuration, when the necessary registered words are different for each small group, the number of registered words in the dictionary can be minimized, so that the word search is also performed at high speed. Furthermore, even if an incorrect word is registered in a small group, there is an effect that the entire large group is not affected.

  Referring to FIG. 3, when the operation of the document check apparatus 100 is started, first, group information for specifying a content field (customer) of document data to be checked is displayed by the user, such as a keyboard 110 and a mouse 112. (S100). In response to this, the document data analysis unit 120.2 uses the partial dictionary 135. i (1 ≦ i ≦ n) is selected.

  Subsequently, the document fetch unit 120.1 fetches the document data 136 and develops it on the RAM, which is a working memory (S102).

  The document data analysis unit 120.2 selects / specifies the description of “claims” to be checked from the document data 136 (S102), and then nouns and noun phrases that are candidates for constituent elements To extract the common dictionary data 134 and the selected partial dictionary 135. Document data is analyzed using i and the morphological analysis engine (S106). Subsequently, the document data analysis unit 120.2 specifies the constituent element candidates based on the analysis result, and the use of the above-described “character”, which is a specific character string, between the constituent element candidates, In the description of the claim, it is checked whether or not they match (S108).

  In accordance with the check result, the data display output control unit 120.3 displays the check result on the display of the document data, for example, by highlighting different colors for the component candidates, i) “The above” “the relevant” is attached and the use is appropriate (the same “candidate of component” has already been mentioned in the claims), ii) “the above” “the relevant” is attached Is not used, and its use is appropriate (the “candidate of component” appears for the first time in the claims), iii) “the above” and “the relevant” are attached, Iv) “Appropriate use” is not displayed, but “appropriate use” is displayed in a distinguishable display mode. Since i) and ii) both display appropriate items, they may be displayed in a common display mode.

  Next, if re-analysis is performed based on an input from the user (S112), the process returns to step S102, and if re-analysis is not performed, the process ends.

Here, the user instructs reanalysis when it is necessary to reanalyze when the user corrects the document data 136 based on the display result in step S110. At this time, as described above, the user corrects a file created by specific word processor software, and the document data 136 to be checked can be configured to take in only a text file from the file. .
[Details of document check processing]
Hereinafter, the document check process described with reference to FIG. 3 will be described in more detail.
(State of existence of component)
First, the condition of the state in which a component exists will be described.
It is assumed that the component exists in the following state in the sentence.
1. Components themselves 1-1 When configured with nouns (example: semiconductor laser element, control device, etc.)
1-2 When configured with a modifier + noun (eg, specific semiconductor laser element, predetermined control device, etc.)
1-3 Consists of technical terms and proper nouns (example: Pn junction)
Therefore, it is not sufficient to extract a constituent element simply by specifying a noun by morphological analysis.
2. Character string behind component 2-1 Component + particle (example: in storage device)
2-2 Constituent element + auxiliary verb (Example: storage device)
2-3 Constituent elements + last word (eg, for each storage device, above the set temperature, between hydrogen electrodes, to the storage device (3))
2-4 Component + Punctuation (eg, storage device, storage device)
2-5 Component + other than above (eg, after the game state ends)
Therefore, conversely, if these “character strings behind the constituent elements” are used as pointers in the document data, they can be used to find end point candidates of the constituent elements.
3. Character string in front of the component 3-1 Other than the noun + component (eg, in a storage device having a ...)
3-2 Noun + Constituent elements (Example: ... in the storage device ...)
3-3 Prefix + component (example: for each storage device ...)
3-4 The above-mentioned + component (eg, .. to the storage device)
The starting point of a component may not be simply broken by a noun.
4). Component substring (special)
4-1 Noun + verb + specific word (Example: selection step)
4-2 Noun + verb + particle + specific word (eg, step for selection)
The specific word is a word registered in a dictionary P2 of a dedicated dictionary described later. That is, the constituent element may be simply a compound word of a verb or the like and a noun other than the connection of nouns.

(Document check processing flow)
FIG. 4 is a flowchart for explaining in more detail the processing in steps S106 and S108 in FIG.

7 and 8 are diagrams for explaining processing in each processing step of FIG.
First, referring to FIG. 4, when the selection of the data range is completed in step S104, the document data analysis unit 120.2 extracts each of the claims separately and extracts the dependency between the claims (S106). 1.1). Such claims can be separated by using tags in the original text data. In addition, the dependency relationship can be identified by extracting an expression such as “described in claim ○ or Δ” by matching with a template using a regular expression.

  Subsequently, the document data analysis unit 120.2 checks whether the claims are serial numbers or whether the dependent claim and the name of its own invention match (S106.1.2). . Although not particularly limited, for example, “name of invention” may be i) specified by the user, or ii) extracted as a constituent element at the rearmost (end) side in the claim. Can be used. In the case of ii), the consistency check of the title of the invention is performed simultaneously with the consistency check such as “above” described later.

  Subsequently, referring to FIG. 4 and FIG. 7, the document data analysis unit 120.2 performs morphological analysis. First, the document data analysis unit 120.2 forces a special part of speech before morphological analysis for a specific word registered in the dedicated dictionary as preprocessing for morphological analysis (referred to as “front-end processing”). Assigning process is performed (S106.2.1).

FIG. 5 is a diagram showing a list of such special parts of speech.
First, parts of speech are assigned in the “front end processing” as follows.

  1) Partial dictionary 135. The user is registered in i (part of speech P0). A character string that is considered a component. However, only when a noun is prefixed, it is considered as a constituent element including it, so it is classified as a “constituent suffix”. Partial dictionary 135 which is a user dictionary. Only this part of speech P0 is registered in i, and the other parts of speech are registered in the common dictionary data 134.

2) Component prefix (parts of speech H1, H2, H3)
2-1) Part of speech H1: A word that comes at the top of a component. A check character string (a specific character string such as “above” or “related”) is not placed after the check character string. For example, “specific”.

    2-2) Part-of-speech H2: A word that comes at the beginning of a component. Check string may be appended. For example, “plurality”, “predetermined”, and the like. In other words, when the “plurality of XX devices” has already appeared, when they are described again later, they may be “the plurality of XX devices” or “the plurality of the XX devices”.

    2-3) Part of speech H3: A word that comes to the top of the component. There may be multiple combinations with conjunctions. For example, “first” + number, “first” + number + “no”, and the like. At this time, it differs from the part of speech H1 or the part of speech H2 in that it is used as “first and second OO”.

3) Suffix (part of speech TT) (this is stored in the dictionary FT in the suffix dictionary)
Regardless of the analysis of the part of speech in the morphological analysis, whenever it is used in the claims, it always follows the component. For example, “(”).

4) compulsory noun (part of speech N) (this is registered in the dictionary FN)
Regardless of the analysis of the part of speech in the morphological analysis, the noun N is always assigned. This is because the description in the claims may be used in a component. For example, “˜”, “/”, and the like.

5) Check character string (part of speech ZZ)
It is a word that precedes a component, and is a word that is checked for consistency in its use as a “specific character string”. The relevant words are “the above”, “the”, “the”, “the above”, “each”, “each”, and the like. The reason why “each” and “each” are used as check character strings is as follows. That is, “each” may be determined as “prefix and noun connection” in the morphological analysis. By the way, in general, for recognition as a constituent element, it is appropriate to treat the word “prefix and noun connection” as a unit with the following noun, but for “each”, “each said ○○” , “Each said XX” (XX is a noun) may be used. In the latter case, according to this principle, “each ○○” is considered as a component, so rather, the exception handling of checking “each” and “each” as a check character string is appropriate. .

  Returning to FIG. 4 and FIG. 7 again, when the front end processing is completed, the document data analysis unit 120.2 performs morphological analysis processing, and nouns for words other than those for which part of speech is forcibly assigned by the front end processing. N, prefix and noun connection N0, conjunction O, particle J, phrase K1, reading K2, verb V, auxiliary verb G, etc. are assigned.

  Next, the document data analysis unit 120.2 uses the result of the morpheme analysis as a condition, and for a specific word registered in the dedicated dictionary as post-processing of the morpheme analysis (referred to as “back-end process”). After the morphological analysis, a process of forcibly assigning a special part of speech is performed (S106.2.3).

  Referring to FIG. 5 again, parts of speech are assigned in the “back end processing” as follows.

1) Constituent Suffix When the word is determined to be a noun by morphological analysis, and [Participant or Auxiliary Verb or Punctuation or Suffix or Conjunction] is postfixed, the common dictionary data 134 If there is a registration as the following part of speech of the component suffix, any of the following parts of speech P1 to P3 is assigned to the word.

    1-1) Part of speech P1: A word that is the end of a component. The verb is not prefixed. For example, “device”, “element”, and the like.

    1-2) Part-of-speech P2: A word that is the end of a component. Verb prefix is allowed. The relevant word is, for example, “step”, “process”, “means”.

  That is, descriptions such as “step to do” and “means to do” are possible for the word of part of speech P2.

    1-3) Part of speech P3: A character string that is regarded as a component only when a noun is prefixed. The relevant word is, for example, “condition”. This is a noun that is assumed to be used as a general name, and when it refers to a specific thing, it is a noun that forms a noun phrase (compound noun) with a noun that precedes it.

2) Suffix (part of speech TT) (this is stored in the dictionary TT in the suffix dictionary)
This is a word that is postfixed to a component when it is determined as a noun by morphological analysis. For example, “every”, “above” or “below”.

3) Non-noun (part of speech XN)
Even if it is determined as a noun by morphological analysis, the recognition is canceled and the part of speech XN is assigned. For example, “Uchi” “Yo”. This is because these words, even if they are nouns, are not appropriate to include in the component. This is a case where it is used in such a manner that “specific XX among a plurality of XX” and “switch processing to do”.

4) Non-prefix (XS)
Even if it is recognized as “prefix and noun connection” in the morphological analysis, the recognition is canceled and the part of speech XS is assigned. For example, “each”. This means that when “multiple XX” has already been mentioned and “each XX” is described, it is originally described as “each XX” or “each XX” in the following description. Is appropriate. However, if “including each” is determined as a constituent element, “each OO” is the first appearance, and therefore may be erroneously determined as appropriate as a usage mode.

  4 and 8, when the back-end process is completed, the document data analysis unit 120.2 performs a correction process (S108.1) as a data recognition process (S108). In the correction process, the document data analysis unit 120.2 assigns an appropriate symbol again when the arrangement of the assigned symbols representing the assigned part of speech satisfies the conditions of other items. That is, when a term that should be recognized as one constituent element is recognized as being separated into a plurality of words, these words are concatenated and an assignment symbol is assigned to the concatenated words. .

  For example, if words designated as noun N are consecutive, these words are concatenated and the symbol of noun N is assigned to the concatenated terms. In addition, when a word recognized as a noun connection N0 is consecutive before a word that is designated as a noun N, these words are concatenated, and the symbol of the noun N is added to the concatenated terms. Assign.

  When the component prefixes H1 and H1 are continuous, these words are concatenated, and the part of speech H1 is assigned to the concatenated words.

  Alternatively, if a word recognized as a verb V is consecutive before a word recognized as a component suffix P2, these words are connected to form part of speech P (component suffix P0, P1). , P2 and P3 are collectively referred to as part of speech P).

  In addition, when the word recognized as the noun N continues before the word of the component suffix P, the document data analysis unit 120.2 combines these words and combines them. Assign part of speech P to

If such a mode of word connection processing is expressed by a part-of-speech symbol, the following cases may occur.

1) N + N → N
2) N0 + N → N
3) H1 + H1 → H1
4) H3 + H1 → H1
5) H3 + H2 → H1
6) V + V → V
7) N + V → V
8) V + P2 → P
9) V + N + J + P2 → P
10) N0 + P1 → P
11) P0 → P
12) P1 → P
13) P2 → P
14) N + P3 → P
15) N + P → P
16) P + P → P
Accordingly, there is no distinction between the component suffixes P0, P1, P2, and P3. After all, a part of speech P that is a single symbol is assigned to a term that should be recognized as one component by correction processing. Become.

  Further, when the correction process is completed, the document data analysis unit 120.2 performs a component element recognition process (S108.2) as a data recognition process (S108).

  In the component recognition process, when the arrangement of the assigned symbols satisfies the conditions of other items in the list shown in FIG. 8, it is recognized as a component. Recognize the [○ + ○ + (...)] part of the “Other” item in the list as a component, and assign a symbol indicating the presence or absence of the check character string.

  Here, FIG. 6 is a diagram showing symbols assigned in the component recognition process in this way.

  The symbol ZC means a component preceded by a check character string (said,..., Etc.), and the symbol C means a component not preceded by a check character string (said, .., etc.). To do.

  For example, when a term of the part of speech P exists after the check character string ZZ, the symbol ZC is assigned with the term (a word or word group to which the symbol [P] is assigned) as a constituent element. On the other hand, when the part of speech P term is not continuous with the check character string, the term C is assigned to the term as a constituent element.

  For example, when the sequence of assigned symbols is ZZ + [H2 + H3 + P], the constituent element (a term consisting of a group of words to which the symbol [H2 + H3 + P] is assigned) is assigned the symbol ZC, while when it is [H2 + H3 + P], A symbol C is assigned to the component (a term consisting of a word group to which the symbol [H2 + H3 + P] is assigned).

Such component recognition processing will be described as follows.
1) ZZ + [H3 + O + H3 + P]: ZC
2) ZZ + V + G + [H3 + O + H3 + P]: ZC
3) [H3 + O3 + H3 + P]: C
4) ZZ + [H2 + H3 + P]: ZC
5) ZZ + V + G + [H2 + H3 + P]: ZC
6) [H2 + H3 + P]: C
7) ZZ + [H3 + P]: ZC
8) ZZ + V + G + [H3 + P]: ZC
9) [H3 + P]: C
10) ZZ + [H2 + P]: ZC
11) ZZ + V + G + [H2 + P]: ZC
12) [H2 + P]: C
13) ZZ + [H1 + P]: ZC
14) ZZ + V + G + [H1 + P]: ZC
15) [H1 + P]: C
16) ZZ + [P]: ZC
17) ZZ + V + G + [P]: ZC
18) [P]: C
Here, the document data analysis unit 120.2 determines that “the first and second OO” is “ZZ + [H3 + O + H3 + P]”, ie, “the first and second OO”. A process is performed assuming that “the second XX” is described. Further, the document data analysis unit 120.2 determines that [H3 + O + H3 + P], that is, “first and second XX”, for example, “first XX” and “second XX”. ”Is performed.

  With the above processing, it is possible to determine whether or not the check character string ZZ such as “above” is prefixed for the terms extracted as candidates for each component, the document data analysis unit 120.2 Subsequently, it is checked whether the use of such a check character string is consistent with the appearance of the component in the claim for the first time or already (S108.3).

  At this time, for each claim, a range for determining whether the claim is an independent claim or a dependent claim is set depending on whether the claim is an independent claim or a dependent claim.

  For example, claim 3 is dependent on claim 2, and when claim 2 is dependent on claim 1, not only claim 3 but also claim of components appearing in claim 3 The first appearance or the previous appearance is determined within the range including Item 2 and Claim 1.

  FIG. 9 shows an example of a dictionary in which terms (not limited to words) corresponding to the part of speech P0 among the component suffixes that are user dictionaries are registered.

  While the common dictionary is registered by the system administrator, in the dictionary P0 (user dictionary), specific terms (words or word groups) are registered by the user according to the business scope of a certain customer. The

  FIG. 10 shows an example of a dictionary in which terms corresponding to the part of speech P1 are registered among component suffixes that are common dictionaries.

  In this dictionary P1, words that are commonly used at the end of the constituent elements are registered regardless of the customer (or the business scope of the customer or the technical field).

  As described above, the dictionary P1 that is a common dictionary is registered by the system administrator. When a common dictionary such as the dictionary P1 and a user dictionary are placed in an external storage device such as a server, and a plurality of terminals share and use the common dictionary and the user dictionary, system management is collectively performed on the server side. The user can register words in the common dictionary, and the user dictionary (user dictionary) can be registered by the user from the client terminal side.

  FIG. 11 shows a dictionary in which terms corresponding to the part of speech P2 and the part of speech P3 are registered among the common dictionaries, a dictionary in which terms corresponding to the parts of speech H1, H2, and H3 are registered, and suffixes. The example of the dictionary which registered the term corresponding to TT is shown.

  In the example shown in FIG. 11, “step”, “process”, and “means” are registered in the part of speech P2, and “condition” is registered in the part of speech P3. Other parts of speech H1, H2, and H3 are as described above.

  FIG. 12 shows an example of a dictionary in which terms corresponding to suffix FT, compulsory noun FN, check character string ZZ, non-noun XN, and non-prefix XS are registered in the common dictionary.

The contents of these words have been described above and will not be repeated.
FIG. 13 is a diagram conceptually illustrating the correction process and the component recognition process performed by the document data analysis unit 120.2.

  For example, since “means” is registered as part of speech P2, the nouns “read” and “control” consecutive before “means” are connected to “means” and recognized as the component “read control means”. Since the check character string is not prefixed, the assignment symbol C is assigned.

  On the other hand, in the “reading balancer”, “balancer” is not registered as part of speech P1, P2, P3, and therefore, it is not recognized as a component when the first component recognition process is completed. become. However, in step S110 of FIG. 3, the user selects, for example, “balancer” as a partial dictionary 135. When registered in i, the “balancer” and the “noun” that is a consecutive noun in front of this are combined, and the “reading balancer” is recognized as one component. Since “check balancer” is preceded by a check character string, an assignment symbol ZC is assigned.

  FIG. 14 is a diagram showing an example of the description of the claim displayed in step S110 of FIG. For example, what is recognized as a component is underlined. Of course, what is recognized as a component may be highlighted with a specific color, for example. Furthermore, since the term “intermediate code” is not recognized as a component, for example, when the user operates the mouse 112 and selects the range of the term “intermediate code”, the partial dictionary 135 . i is registered.

FIG. 15 is a diagram showing an example of the entire screen displayed in step S110 of FIG.
In the upper left, the dependency relationship of the detected claims is displayed as a claim tree. In the upper right, a list of components is displayed.

  The detected “component” is underlined or highlighted for each claim.

  At the same time, “component name” detected for each claim is displayed as a list with check boxes. In the dependent claims, only newly detected items in the tree are displayed as a list. The check box is an interface for deleting unnecessary items from “component candidates”.

  As described above, the color of the highlight display is changed depending on whether or not the check character string is properly used, whether or not the description such as “said” is unnecessary, or whether or not the description such as “said” is missing. Thus, it is possible to inform the user of the consistency state of the use of the check character string.

  As described with reference to FIG. 14, when the user confirms a term that is not recognized as a component at this display stage, the user selects a range of the term, whereby the partial dictionary 135. The term is registered in i. At that stage, when the user clicks the “Analyze” button, the updated partial dictionary 135. Based on i, the component is recognized and the consistency is checked as described above.

  With the configuration as described above, in a document to be analyzed, a specific character string is used in order to indicate that a term following the specific character string has already appeared in the document. It is possible to easily check whether it is appropriate.

  In addition, even if the term is not registered in the user dictionary one by one, the specific term is identified using information on multiple term suffixes commonly used at the end of the term, so the user identifies the specific term It is possible to check the document by greatly reducing the processing to be performed.

[Embodiment 2]
When document check apparatus 100 according to Embodiment 1 is used to indicate that a specific character string in a document data given from the outside has already appeared in the document, the term following the character string In addition, the description has been made assuming that a word / phrase that is a candidate for a term is extracted from a document, and whether or not the use of such a specific character string is consistent among the extracted candidates. In this case, the document checking apparatus 100 according to the first embodiment has been described as performing a document check using “claims” as a check target.

  At this time, with regard to the above term (ie, “specific term”), a candidate for a specific term is extracted from the sentence by using the high probability that the end of the term will be a fixed character string (“term suffix”). Used as a clue. Here, when the sentence to be checked is a claim, the “specific term” is a “component” and the “term suffix” is a “component suffix” .

However, the following cases can be considered as the document check.
1) When extracting specific term candidates, if nouns identified by the morphological analysis engine are consecutive without using "term suffix" as a clue, the specific terms can be linked by concatenating these nouns. The same processing can be realized even in a configuration in which candidates that are inappropriate as “specific term candidates” are excluded by a user's individual instruction as necessary.

  In particular, when the document to be checked is “specification”, a character string (for example, “means”) such as “component suffix” is not necessarily used, unlike “claim”. In this case, there may be cases where noun linking is more suitable for extracting candidates for “explanatory terms” described later.

  2) Also, depending on the document to be checked, in a part of the document, as described above, a specific character string may indicate that a term following the character string has already appeared in the document. In some cases, the concept is defined while ensuring the strictness of the use of the term, and the concept is specifically explained in other parts of the document. In this case, the specific example as described above is explained by a term indicating a more specific example (hereinafter referred to as “explanatory term”) corresponding to the “specific term” used in the definition of the concept. May be. At this time, in the specific description as described above, when explanation is made with reference to the drawings, such explanation terms are immediately followed by reference numerals in the explanation to indicate correspondence with the drawings. In the corresponding drawings, the same reference numerals are given to the corresponding portions of the drawings, and a specific description may be executed.

  For example, when “definition of concept” is described in “claims”, such specific explanation is given by “specification” (particularly, for example, “implementing the invention”). The description of "form for doing"). At this time, the “descriptive term” is a more specific explanation of the “specific term”, and therefore, either the same or similar word or combination of words is explicitly used for both, Alternatively, in the common sense of the reader who reads the sentence, an explanatory term that makes it obvious that such an “explanatory term” corresponds to a “specific term” is used.

  At this time, since “explanatory terms” are immediately followed by a reference numeral as described above, the correspondence between the explanatory terms and the corresponding reference numerals is consistent in the document. Therefore, it is necessary to check whether they are consistent.

  And, as described above, the part-of-speech identification by the “morpheme analysis engine” includes a part where such “specific term” is used and a part where “explanatory term” is used in the text to be checked. It becomes possible to carry out in common.

  For example, in patent application documents, “claims” and “specifications” are different sentences in form and may be created as separate data files. However, in the present specification, when both have a close relationship such as “definition of concept” and “specific explanation of the concept” as described above, This will be called “check target document”.

  Hereinafter, the document check apparatus, the document check method, and the document check program according to the second embodiment will be described based on the above premise. Here, the hardware configuration and the functional configuration of the document check apparatus according to the second embodiment are basically the same as those of the document check apparatus 100 according to the first embodiment described with reference to FIGS. 1 and 2. .

Hereinafter, the operation of the document checking apparatus according to the second embodiment will be described.
In the following description of the operation of the document checking apparatus according to the second embodiment, an example in which “claims” and “specifications” are set as documents to be checked will be described.

  Here, in the following description of the second embodiment, the term “component” is a generic term for both “specific term” in “claims” and “explanatory term” in “specification”. Use as a term.

  FIG. 16 is a flowchart for explaining the operation of the document check apparatus according to the second embodiment, and is a diagram contrasted with FIG. 3 according to the first embodiment.

  Referring to FIG. 16, when the operation of the document check apparatus 100 is started, first, the group information for specifying the content field (customer) of the document data to be checked by the user is the keyboard 110, the mouse 112, and the like. (S200). In response to this, the document data analysis unit 120.2 uses the partial dictionary 135. i (1 ≦ i ≦ n) is selected.

  At this time, it is also possible to selectively set whether or not to use the “common dictionary” by user input.

  Here, it is assumed that the contents of the common dictionary and the customer dictionary are stored in correspondence with “claims” and “specifications”, respectively. However, for example, common dictionary data and customer dictionary data corresponding to “Claims” are stored in a storage device of another computer connected via a network, and the document check device 100 includes a communication interface 128. These data are accessed and read out via the URL, or written, while the common dictionary data and customer dictionary data corresponding to the “specification” are stored in the hard disk 124 in the document checking apparatus 100. It is also possible to adopt a configuration. This is due to the use of terms that are relatively uncommon in documents that make “concept definitions” such as “claims”, and as a result of use by multiple users, for example, This is because by sharing the data registered in the dictionary data among a plurality of users, other users can use the registration by other users, that is, a kind of learning result. On the other hand, because the term “specification” tends to use commonly used terms, it is not always necessary to share the results of such learning compared to the claims. is there.

  Subsequently, the document fetch unit 120.1 fetches the document data 136 and develops it on the RAM, which is a working memory (S202).

  The document data analysis unit 120.2 selects and identifies the description of “claims” and “specifications” to be checked from the document data 136, and selects nouns and noun phrases that are candidate components. In order to extract the common dictionary data 134 and the selected partial dictionary 135. The document data is analyzed using i and the morphological analysis engine, and a symbol as a pre-process for specifying the part of speech is assigned to a part of the character string of the document data in accordance with the registered contents of the dictionary (S206). Subsequently, the document data analysis unit 120.2 assigns a symbol for specifying the part of speech to each character string based on the analysis result in the document data according to the part of speech specified by the morphological analysis (S206). .

  Subsequently, the document data analysis unit 120.2 has a character string registered in the dictionary and excluded from the recognition target as a noun for the character string to which the symbol is assigned by the above processing, For the character strings registered in the dictionary after the element, the character strings registered as nouns (component suffixes) at the end of the component, etc. Perform (S208). After that, the document data analysis unit 120.2 concatenates the assigned symbols according to a predetermined rule, and reallocates them by integrating them (S208). Further, the document data analysis unit 120.2 makes a “specific character string” indicating that the component has already been issued, that is, within the scope of the claims, according to the presence or absence of “the above”. Symbols for recognizing consistency (ZC, C: description will be described later) are assigned. Thereby, the candidate of a component is specified. However, when “the above” is not used in the specification, such a “symbol for recognizing consistency” does not need to be distinguished in the specification in the determination of consistency. It only indicates that it is a “candidate”.

  Subsequently, the document data analysis unit 120.2 checks whether the use of the specific character string “said” “related” is consistent among the candidate components in the claims. (S210). Further, the document data analysis unit 120.2 checks whether the correspondence between the component and the code is consistent in the specification (S210).

  In accordance with the check result, the data display output control unit 120.3 displays the check result on the display of the document data, for example, highlights of different colors for the candidate components in the claims. I) “the above” and “appropriate” are attached and the use thereof is appropriate (the same “candidate of component” has already been mentioned in the claims), ii) “ The above-mentioned "" concerned "is not attached, and its use is appropriate (the" component candidate "appears for the first time in the claims), iii)" the "" related "is attached Although not used, iv) “the above” and “not applicable” but not used are displayed in a distinguishable display mode (S212). Since i) and ii) both display appropriate items, they may be displayed in a common display mode.

  Further, the data display output control unit 120.3 has a one-to-one relationship between the component and the sign for the specification according to the check result (the use is likely to be correct). Ii) One component using a plurality of symbols (possibly misused), iii) One component using a plurality of components (clearly misused), Each of them is displayed with distinction according to a display mode such as a color. Thereby, the user can recognize the component that is clearly misused and the component that may be misused in the specification. It should be noted that the relationship between the component and the code is a one-to-one relationship, but “use is likely to be correct” is considered to be the use of the component and the code in the document in the above check. This is because no check is made up to the consistency with the use of symbols in the drawing. However, if the drawing is also electronic data, such consistency may be checked. For example, in the description part of FIG. X, it may be checked whether the code of the component used is also present in the corresponding drawing data of FIG.

  Next, if reanalysis is performed based on an input from the user (S214), the process returns to step S202. If no reanalysis is performed, the process ends.

Here, the user instructs reanalysis when it is necessary to reanalyze when the user corrects the document data 136 based on the display result in step S212. At this time, as described above, the user corrects a file created by specific word processor software, and the document data 136 to be checked can be configured to take in only a text file from the file. .
[Details of document check processing]
Hereinafter, the document check process described with reference to FIG. 16 will be described in more detail.

First, FIG. 19 is a diagram showing the configuration of the dictionary according to the first embodiment and the second embodiment in comparison.
As shown in FIG. 19, in the second embodiment, the dictionary name and the contents of the dictionary are changed from the first embodiment.

  However, most dictionaries are common to the first embodiment. The default word registration of all dictionaries can be the same in the first embodiment and the second embodiment.

  However, although the H3 dictionary can be the same as that of the first embodiment, it can be generated internally using a regular expression and can be configured not to be provided as a dictionary. The FT and XS dictionaries can be the same as those in the first embodiment. However, it is also possible to change the symbol assignment in the morphological analysis so that the same processing result is obtained. Furthermore, the P1 dictionary is configured to be able to select whether to use or not. Also, the TNC recognition method is changed, and the type of dictionary is changed to “customer dictionary”.

  Further, FIG. 20 is a diagram illustrating “assignment symbols” assigned to character strings in a document to be checked.

The meaning of “assignment symbol” is as follows.
ZC: component preceded by the above C: component not preceded by the above P: component candidate H: component prefix string candidate XX: inappropriate component The component recognition process is as follows. The target character string is divided using the dictionary or morphological analysis information, and a symbol is assigned to the divided character string. In the following description, the assigned symbol is “assignment symbol”, and the assignment condition is “assignment condition”.

(Document check processing flow)
FIG. 17 is a flowchart for explaining the processing of step 206 in FIG.

  Referring to FIG. 17, in step S206.1.1, it is determined whether or not the processing has been completed for all of the acquired documents. If completed, the process proceeds to step S210. If not completed, the process proceeds to the next step S206.1.2, and a document to be processed is specified. For example, if the text data to be processed is configured to include “specification data” after “claimed data”, first, “claimed data” becomes the processing target.

  Subsequently, if the processing target is the scope of claims (step S206.1.3), the document data analysis unit 120.2 extracts the claims separately for each claim and extracts the dependency relationship between the claims. (S106.1.4). Such claims can be separated by using tags in the original text data. In addition, the dependency relationship can be identified by extracting an expression such as “described in claim ○ or Δ” by matching with a template using a regular expression.

  Subsequently, the document data analysis unit 120.2 checks whether the claims are serial numbers or whether the dependent claim and the name of its own invention match (S106.1.5). . Although not particularly limited, for example, “name of invention” may be i) specified by the user, or ii) extracted as a constituent element at the rearmost (end) side in the claim. Can be used. In the case of ii), the consistency check of the title of the invention is performed simultaneously with the consistency check such as “above” described later.

  Subsequently, the document data analysis unit 120.2 performs front end processing (S206.2.1).

  Here, FIG. 21 is a diagram for explaining the assignment symbols assigned in the front-end process.

  As shown in FIG. 21, symbols are assigned to the target character string according to the rules in FIG. 21, based on the dictionary shown in FIG. The allocation process is performed according to the order in the table. When the order is described in the table, the same applies to other processes.

  Subsequently, returning to FIG. 17, the document data analysis unit 120.2 executes morpheme analysis (S206.2.2).

FIG. 22 is a diagram for explaining assignment symbols in such morphological analysis.
As shown in FIG. 22, morpheme analysis processing is performed on a character string to which no symbol is assigned in the front end processing. In the morphological analysis process, a character string obtained by dividing the input character string and a part of speech corresponding to the character string are obtained. According to the table of FIG. 22, symbols are assigned to the divided character strings based on the part of speech (noun: N, conjunction: O, verb: V, auxiliary verb: G, particle: J).

  FIG. 18 is a flowchart for explaining in detail the flow of steps S208 and S210 of FIG.

  In step S208, the document data analysis unit 120.2 first performs backend processing 1 (S208.1).

  FIG. 23 is a table for explaining such back-end processing 1. As shown in FIG. 23, in the back-end process 1, symbols are assigned according to the table of FIG. 23 based on the dictionary with respect to the character strings to which symbols have already been assigned in the process so far.

  Returning to FIG. 18, next, the document data analysis unit 120.2 performs correction processing 1 to perform symbol replacement and concatenation processing (S208.2).

  FIG. 24 is a table for explaining such correction processing 1. As shown in FIG. 24, in the correction process 1, symbols are assigned according to the table of FIG. 24 based on the dictionary with respect to the character strings to which the symbols have already been assigned in the process so far.

  For example, an assigned symbol P is assigned to a character string registered in the TOC dictionary, and an assigned symbol N is assigned to a character string registered in the TON dictionary. Further, in the same manner as in the first embodiment, a process of connecting nouns (N + N → N) is performed.

  Returning to FIG. 18, next, the document data analysis unit 120.2 registers a noun or a connected noun as a constituent element candidate when the P1 dictionary is not used (S208.3). Correction processing 2 is performed (S208.4).

FIG. 25 is a table for explaining such correction processing 2.
Returning to FIG. 18, next, the document data analysis unit 120.2 performs a component recognition process (S208.5).

FIG. 26 is a table for explaining such a “component recognition process”.
As shown in FIG. 26, symbols are assigned according to the rules in the table of FIG. 26 to the character strings to which symbols have already been assigned in the above processing. However, a symbol is assigned to the character string corresponding to the part enclosed in square brackets. For example, taking the order 1 in the table of FIG. 26 as an example, the character string corresponding to “H + P” is assigned “ZC” because the character string ZZ exists immediately before that. Taking order 3 as an example, a character string corresponding to “H + P” is assigned “C” because there is no character string ZZ immediately before it.

  Returning to FIG. 18, next, when the processing target is “specification” (S208.6), the document data analysis unit 120.2 performs back-end processing 2 (S208.7).

  FIG. 27 is a table for explaining such back-end processing 2. If the character string to which the symbol “C” or “ZC” is assigned is an alphanumeric character string, that is, if the code is not followed, the symbol XX is assigned.

  Returning to FIG. 18, next, the document data analysis unit 120.2 performs the back-end process 3 according to the registered contents of the TNC dictionary (S208.8).

  FIG. 28 is a table for explaining such back-end processing 3. If the corresponding word is registered in the TNC dictionary for the character string assigned the symbol “C” or “ZC”, the symbol XX is assigned.

  Returning to FIG. 18, the document data analyzing unit 120.2 then checks the consistency of “claimed” or the like for the claims when the processing has been completed for all the acquired documents (S206.1.1). Is performed in the same manner as in the first embodiment (S210.1).

  Next, the document data analysis unit 120.2 checks the consistency between the component and the reference code for the specification (S210.2).

In this case, the following states are determined for the components in the specification as described above.
State i) The relationship between the component and the sign is a one-to-one relationship (use is likely to be correct),
State ii) A case where a plurality of codes are used for one component (possibly misuse),
State iii) A case where a plurality of components are used for one code (clearly misused).

  FIG. 29 is a diagram illustrating a display example performed by the data display analysis unit 120.3 in step 212.

  In FIG. 29, the component corresponding to the state 2 is underlined in the component corresponding to the state ii, and the component corresponding to the state iii is surrounded by a frame.

  Of course, the states i to iii may be distinguished by performing highlight processing with different colors.

  With the configuration as described above, “specific term”, “specific character string” indicating that the specific term has already appeared, “descriptive term”, and “sign” attached to the explanatory term are used. It is possible to check the consistency of each use in the document.

  In addition, when a candidate for a specific term is extracted, the use of a “term suffix” as a clue can be selected by a user setting.

  The embodiment disclosed this time should be considered as illustrative in all points and not restrictive. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims.

  100 Document Check Device, 102 Computer Main Body, 104 Monitor, 105 Bus, 106 FD Drive, 108 Optical Disk Drive, 110 Keyboard, 112 Mouse, 122 Memory, 124 Hard Disk, 128 Communication Interface, 131 Display Control Program, 132 Document Data Analysis Program, 133 Group information, 134 common dictionary data, 135 customer dictionary data, 136 document data.

Claims (20)

A document check device for checking consistency of descriptions in a document to be analyzed,
Storage means for storing document data representing the document;
Information acquisition means for acquiring information on the specific character string when a specific character string is used to indicate that a term following the specific character string has already appeared in the document;
A part of speech specifying means for specifying a part of speech included in the document,
The part of speech specifying means is:
A specifying unit for recognizing and specifying the specific character string based on the information acquired by the information acquisition unit in the document;
Morphological analysis is performed on the document, and morphological analysis means for specifying a part of speech in the document includes:
Based on the result of the morphological analysis, term candidate recognition means for identifying a term candidate that is a candidate for the term by connecting consecutive nouns in the document;
Consistency checking means for checking consistency of use of the specific character string between the specified term candidates;
A document check device further comprising display control means for causing the display device to display the consistency check result. The information acquisition means further acquires information of a plurality of term suffixes commonly used for the ending of the term,
The term candidate recognizing means, based on the result of the morphological analysis, connects a noun that precedes the term suffix in the document to the term suffix, thereby identifying a term candidate that is a candidate for the term. The document checking apparatus according to claim 1, wherein the document checking apparatus is specified. The information acquisition unit further registers the term candidates that have not been specified at the time of the process of specifying the term candidates in the term candidate recognition unit in the document as the term candidates by user selection. Get user dictionary information,
The document check apparatus according to claim 2, wherein the term candidate recognition unit further identifies the term candidate with reference to the user dictionary. The storage means stores information on the specific character string, information on the plurality of term suffixes, and the user dictionary,
The document check apparatus according to claim 3, wherein the information acquisition unit reads information on the specific character string, information on the plurality of term suffixes, and information on the user dictionary from the storage unit. The documents are classified into a plurality of groups for each content field,
The said user dictionary is divided | segmented into the partial dictionary for every said group, The said user registers the said term candidate which was not specified to the said partial dictionary corresponding to the said content field of the said document. The document check device described. The storage means further stores in advance a specific prefix prefixed to the term,
The term candidate recognizing means, after concatenating a noun consecutive before the term suffix to the term suffix, and when the specific prefix continues before the term after the concatenation, The document checking apparatus according to claim 4, wherein the term candidates are specified by further connecting words. The specific character string information, the plurality of term suffix information, and the user dictionary are stored in an external storage device outside the document check device,
The document according to claim 3, wherein the information acquisition unit acquires, from the external storage device, information on the specific character string, information on the plurality of term suffixes, and information on the user dictionary by communication. Check device. The document data is
First partial document data representing a first document used by the specific character string to indicate that a term following the specific character string has already occurred in the document;
A second partial document for explaining a content defined by the first partial document data, and expressing a second document used by adding a description term corresponding to the term to the reference Data and
The morpheme analysis means performs the morpheme analysis in common with the first and second partial document data,
The document checking apparatus according to claim 1, further comprising a code checking unit for checking the consistency between the explanation term and the code for the second partial document data. A document check program for causing a computer including an arithmetic device and a storage device that stores document data representing the document to execute consistency check of a description in a document to be analyzed,
A step of obtaining information on the specific character string when the arithmetic device is used to indicate that the specific character string is a term subsequent to the specific character string; When,
Identifying a part of speech included in the document,
The step of specifying the part of speech includes
The arithmetic device recognizing and specifying the specific character string based on the acquired information in the document;
The arithmetic unit performs morphological analysis on the document and specifies parts of speech in the document;
The arithmetic device, based on the result of the morphological analysis, identifying consecutive term nouns in the document by connecting consecutive nouns; and
The computing device checking the consistency of use of the specific character string between the specified term candidates;
A document check program for causing a computer to execute a document check process, further comprising the step of causing the calculation device to display the consistency check result on a display device. Obtaining the information includes obtaining information of a plurality of term suffixes commonly used for the ending of the term;
The step of specifying the term candidates is based on the result of the morphological analysis. In the document, a noun that precedes the term suffix is connected to the term suffix in the document, thereby being the term candidate. The document check program according to claim 9, comprising a step of specifying term candidates. In the obtaining step, the term candidates that are not identified at the time of the process of identifying the term candidates in the step of identifying the term candidates in the document are further selected as the term candidates by user selection. Including the step of obtaining information of the registered user dictionary,
The document check program according to claim 10, wherein the step of specifying the term candidates further includes a step of specifying the term candidates with reference to the user dictionary. The storage device stores information on the specific character string, information on the plurality of term suffixes, and the user dictionary,
The document check according to claim 11, wherein the obtaining step includes a step of reading out information on the specific character string, information on the plurality of term suffixes, and information on the user dictionary from the storage device. program. The documents are classified into a plurality of groups for each content field,
The user dictionary is divided into partial dictionaries for each group, and the user registers the term candidates that are not specified in the partial dictionary corresponding to the content field of the document. The document check program described. The storage device further stores in advance a specific prefix prefixed to the term;
The step of identifying the term candidate includes the step of identifying the noun when the specific prefix continues before the term after the concatenation of the noun that precedes the term suffix to the term suffix. The document check program according to claim 12, further comprising the step of specifying the term candidates by further concatenating the prefixes. The specific character string information, the plurality of term suffix information, and the user dictionary are stored in an external storage device outside the computer on which the document check program is executed,
The obtaining step includes a step of obtaining from the external storage device information on the specific character string, information on the plurality of term suffixes, and information on the user dictionary by communication. The document check program described. The document data is
First partial document data representing a first document used by the specific character string to indicate that a term following the specific character string has already occurred in the document;
A second partial document for explaining a content defined by the first partial document data, and expressing a second document used by adding a description term corresponding to the term to the reference Data and
The step of specifying the part of speech includes the step of performing the morphological analysis in common with the first and second partial document data;
The document check process includes:
11. The document check program according to claim 9, further comprising a step of checking consistency between the descriptive term and the code for the second partial document data. A document check method for causing a computer including an arithmetic device and a storage device that stores document data representing the document to execute consistency check of a description in a document to be analyzed,
A step of obtaining information on the specific character string when the arithmetic device is used to indicate that the specific character string is a term subsequent to the specific character string; When,
Identifying a part of speech included in the document,
The step of specifying the part of speech includes
The arithmetic device recognizing and specifying the specific character string based on the acquired information in the document;
The arithmetic unit performs morphological analysis on the document and specifies parts of speech in the document;
The arithmetic device, based on the result of the morphological analysis, identifying consecutive term nouns in the document by connecting consecutive nouns; and
The computing device checking the consistency of use of the specific character string between the specified term candidates;
And a step of causing the arithmetic device to display the consistency check result on a display device. Obtaining the information includes obtaining information of a plurality of term suffixes commonly used for the ending of the term;
The step of specifying the term candidates is based on the result of the morphological analysis. In the document, a noun that precedes the term suffix is connected to the term suffix in the document, thereby being the term candidate. The document checking method according to claim 17, further comprising a step of identifying term candidates. In the obtaining step, the term candidates that are not identified at the time of the process of identifying the term candidates in the step of identifying the term candidates in the document are further selected as the term candidates by user selection. Including the step of obtaining information of the registered user dictionary,
The document checking method according to claim 17 or 18, wherein the step of specifying the term candidates further includes a step of specifying the term candidates with reference to the user dictionary. The document data is
First partial document data representing a first document used by the specific character string to indicate that a term following the specific character string has already occurred in the document;
A second partial document for explaining a content defined by the first partial document data, and expressing a second document used by adding a description term corresponding to the term to the reference Data and
The step of specifying the part of speech includes the step of performing the morphological analysis in common with the first and second partial document data;
The document checking method according to claim 17 or 18, further comprising a step of checking the consistency between the explanation terminology and the code for the second partial document data.

Images