JP2013254366A - Information processing device and related word determination method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an information processing device and a related word determination method capable of appropriately extracting a related word corresponding to a prescribed proper noun, taking a prescribed time width into consideration.SOLUTION: A document information acquisition unit 101 acquires document information with which a document and time information are associated, a morpheme analysis unit 103 analyzes the acquired document information, and a candidate determination unit 106 determines a related word candidate of the place name included in the document information. A totaling unit 107 totals the number of documents including the related word candidate and the place name in the document information generated in a prescribed time width. A determination unit 108 determines whether or not the related word candidate is a related word of the place name by using the totaling result and determination index information which is index information for determining whether or not the related word candidate is a general word, and if it is a related word, a related word registration unit 109 registers information which associates the related word of the place name and the place name.

Description

  The present invention relates to an information processing apparatus for acquiring a related word of a proper name and a related word determination method.

  In recent Web technologies, services such as blogs, microblogs, and SNS, which are used by general users to post sentences, have become widespread. In particular, microblogging is highly real-time in information, and it is expected to extract related words that are words having a high degree of relevance of a certain unique name in a certain time span by using text mining technology.

  In order to extract a word related to a certain word, there is a technique for calculating the degree of association between words (for example, Patent Document 1). In Patent Document 1, not only the relationship between words but also the importance of a word is used for the purpose of obtaining useful information during a search by adding a word related to a query word during the information search. Degrees (word familiarity, topic level, word content depth) are calculated, and related words are extracted. Further, Non-Patent Document 1 shows a method for calculating the degree of association using word co-occurrence information.

JP 2010-244339 A

Shinichiro Ishikawa “How to measure the strength of collocation”

  In the method of Patent Document 1, since related words are extracted on the premise of having a static Web page link structure, there is a problem that related words cannot be extracted from a document generated in real time. is there. Further, Non-Patent Document 1 does not consider extracting related words by paying attention to proper names.

  Therefore, in the present invention, in order to solve the above-described problem, an information processing apparatus and a related word determination method capable of appropriately extracting related words corresponding to a predetermined unique name in consideration of a predetermined time width The purpose is to provide.

  In order to solve the above-described problem, an information processing apparatus according to the present invention includes a document information acquisition unit that acquires document information associated with a document and time information, and a document in the document information acquired by the document information acquisition unit. And a related word candidate determining means for determining a related word candidate that is a candidate of a related word in the proper name included in the document, and the related word candidate among the document information generated in a predetermined time width And a counting unit that counts the number of documents including the unique name, a counting result by the counting unit, and determination index information that is index information for determining whether or not the related word candidate is a general word And determining means for determining whether or not the related word candidate is a related word of the unique name, and, as a result of the determination by the determining means, if the related word candidate is a related word, , And the proper name It comprises registration means for registering with the information.

  The information processing apparatus according to the present invention includes a document information acquisition step for acquiring document information associated with a document and time information, and analyzes a document in the document information acquired in the document information acquisition step. A related word candidate determining step for determining a related word candidate that is a candidate for a related word in the specific name included in the document, and a document including the related word candidate and the specific name among the document information generated in a predetermined time width The related word candidate using a counting step for counting the number of words, a counting result obtained by the counting step, and determination index information that is index information for determining whether or not the related word candidate is a general word Determining whether or not is a related word of the proper name, and, as a result of the determination in the determining step, if it is a related word, the related word of the proper name and the proper name And a, a registration step of registering information that associates.

  According to the present invention, the document information associated with the time information is acquired, the acquired document information is analyzed, and the related word candidate in the proper name included in the document information is determined. The number of documents containing related word candidates and unique names in the document information generated in a predetermined time span is counted, and the result of the counting and an index for determining whether the related word candidate is a general word Using the determination index information that is information, it is determined whether or not the related word candidate is a related word of a unique name. If the related word candidate is a related word, information that associates the related word of the specific name with the specific name Register. This makes it possible to extract an appropriate related word having a proper name. That is, it is possible to register a related word corresponding to the unique name in a predetermined time width by removing the general word. In addition, as a proper name here, a place name, a person name, a movie title, etc. correspond.

  In the information processing apparatus of the present invention, an upper item specifying unit that specifies an upper item of the unique name corresponding to the unique name included in the document in the document information and associates the document information with the upper item. In addition, the determination unit uses, as at least the determination index information, the number of documents including the related word candidates in the document information in which the upper item corresponding to the unique name is linked in a predetermined time span.

  According to this invention, the upper item of the unique name included in the document in the document information is specified, the document information and the upper item are linked, and the upper item corresponding to the unique name in the predetermined time width is By using the number of documents including related word candidates among the associated document information as determination index information, the document information including related word candidates among the document information associated with the higher-level item corresponding to the unique name Based on the number, it is determined whether or not the related word candidate is a related word of the proper name, so that the related word of the proper name can be appropriately determined. Note that the upper item here refers to items belonging to a predetermined unique name in a higher-level concept (for example, an area for a place name or a genre for a movie title).

  In the information processing apparatus according to the present invention, the determination unit uses, as at least determination index information, comparison information between the total number of documents and the number of documents including the related word candidates in a predetermined time width.

  According to this invention, it is possible to determine whether or not a related word candidate is a general word by referring to comparison information between the total number of documents and the number of documents including related word candidates in a predetermined time width. Therefore, it is possible to appropriately determine the related word of the proper name.

  In the information processing apparatus of the present invention, the determination unit uses, as at least the determination index information, general word list information generated based on the appearance degree of the document information in the document.

  According to the present invention, the number of documents including place names and related word candidates in a predetermined time width is counted, and the general word list information generated based on the appearance degree of the document information in the document is used as the determination index information. Since it is judged whether a related word candidate is a general word, the related word of a proper name can be judged appropriately.

  According to the present invention, an appropriate related word having a proper name can be registered. That is, it is possible to appropriately extract related words corresponding to a predetermined unique name in consideration of a predetermined time width.

It is a block diagram which shows the function of the information processing apparatus 100 of 1st Embodiment and 2nd Embodiment. 2 is a hardware configuration diagram of the information processing apparatus 100. FIG. FIG. 3 is a diagram showing a structure of data held in a document information storage unit 102. FIG. 3 is a diagram illustrating a structure of data held in a word storage unit 105. It is a figure which shows the structure of the data currently hold | maintained at the related word memory | storage part 110. FIG. It is a flowchart which shows the process of the information processing apparatus in 1st Embodiment. It is a flowchart which shows the process of the information processing apparatus in 2nd Embodiment. It is a block diagram which shows the function of information processing apparatus 100B of 3rd Embodiment. It is a flowchart which shows the process of the information processing apparatus in 3rd Embodiment.

  Embodiments of the present invention will be described with reference to the accompanying drawings. When possible, the same parts are denoted by the same reference numerals, and redundant description is omitted.

(First embodiment)
FIG. 1 is a block diagram illustrating functions of the information processing apparatus 100 according to the first embodiment. As shown in FIG. 1, the information processing apparatus 100 includes a document information acquisition unit 101 (document information acquisition unit), a document information storage unit 102, a morpheme analysis unit 103 (related word candidate determination unit), and an area addition unit 104 (upper level). Item specifying unit), word storage unit 105, candidate determining unit 106 (related word candidate determining unit), totaling unit 107 (totaling unit), determining unit 108 (determining unit), related word registering unit 109 (registering unit), related word The storage unit 110 is included. In the first to third embodiments, a method of extracting related words for place names as unique names will be described in detail.

  FIG. 2 is a hardware configuration diagram of the information processing apparatus 100. As shown in FIG. 2, the information processing apparatus 100 shown in FIG. 1 physically includes a CPU 11, a RAM 12 and a ROM 13 that are main storage devices, an input device 14 such as a keyboard and a mouse that are input devices, a display, and the like. The computer system includes an output device 15, a communication module 16 that is a data transmission / reception device such as a network card, an auxiliary storage device 17 such as a hard disk or a semiconductor memory, and the like. Each function in FIG. 1 operates the input device 14, the output device 15, and the communication module 16 under the control of the CPU 11 by loading predetermined computer software on the hardware such as the CPU 11 and the RAM 12 shown in FIG. 2. In addition, it is realized by reading and writing data in the RAM 12 and the auxiliary storage device 17. Hereinafter, each functional block will be described based on the functional blocks shown in FIG.

  The document information acquisition unit 101 acquires document information on the Web via a network. Specifically, the document information is acquired from the Web server by HTTP (HyperText Transfer Protocol) communication. Specific examples of document information include blogs and microblogs. The document information acquired by the document information acquisition unit 101 is stored in the document storage unit 102.

  The document information storage unit 102 stores document information. FIG. 3 shows an example of document information. The document information includes at least a document ID, document time information, and a document text. The document ID is an ID for identifying original information acquired from the Internet. For example, the URL of a site corresponds. The time information is the time when the document is created. The text of the document is the content of the posted article.

  The morphological analysis unit 103 extracts a plurality of words by performing morphological analysis on the document information document stored in the document information storage unit 102. Note that the morphological analysis technique itself is well known, for example, as disclosed in Japanese Patent Application Laid-Open No. 7-311769, and thus the description thereof is omitted here.

  The area assigning unit 104 specifies a word corresponding to the place name from the result of analysis by the morphological analysis unit 103, and specifies an area (area ID, area name, etc.) from the place name. As a method for specifying an area, there are a method using correspondence information between an area and a place name (not shown), and a method of inquiring an external computer. The word obtained by the morphological analysis is stored in the word storage unit 105 in association with the document ID and the area specified by the area adding unit 104. The area here corresponds to the upper item.

  The word storage unit 105 stores word information that is information that associates the word extracted by the morphological analysis by the morphological analysis unit 103, the document ID, and the area specified by the area adding unit 104. FIG. 4 shows an example of word information. The word information includes at least a document ID, document time information, a word analyzed by the morphological analysis unit 103, and an area. Further, the information processing apparatus 100 refers to the information stored in the word storage unit 105, extracts a place name, and separately holds the extracted place name as a reference place name group. It is assumed that the reference place name group is de-duplicated. Then, the information processing apparatus 100 refers to the information stored in the word storage unit 105 and separately holds a word group other than the place name as a related word candidate group. It is assumed that the reference place name group is de-duplicated. Further, when a plurality of place names are included in one document information, each may be treated as a place name.

  The candidate determination unit 106 determines one or a plurality of related word candidates that are candidates for the related word of the place name included in the document of the document information from the related word candidate group. Specifically, the candidate determination unit 106 sets words other than the place name in the document including the place name as related word candidates. Note that the candidate determination unit 106 does not treat each word in a sentence not including a place name as a related word candidate.

  The counting unit 107 counts the number of documents that meet a predetermined condition. In the first embodiment, for document information created in a predetermined time width (for example, three days), the number of documents including the place name is counted for a document set in an area to which the predetermined place name belongs (hereinafter, the document name is included). The number of documents is also called “document number A”). In addition, the counting unit 107 counts the number of documents including the word determined by the candidate determining unit 106 for the document set in the area to which the place name belongs in the time span (hereinafter, the number of documents is referred to as “number of documents”). B "). Further, the counting unit 107 calculates the number of documents including both the place name and the word determined by the candidate determination unit 106 for the document information created in the time span, for the document set in the area to which the place name belongs. The number of documents is counted (hereinafter, the number of documents is also referred to as “document number C”).

The determination unit 108 determines whether each related word candidate is a related word of a place name. In the first embodiment, a dice coefficient is calculated based on the following formula (1) using the number of documents A, the number of documents B, and the number of documents C, which are the counting results counted by the counting unit 107, and the calculation is performed. If the result is equal to or greater than the threshold value, it is determined as a related word of the place name, and the related word registration unit 109 registers information that associates the place name with the related word candidate. On the other hand, if the calculation result of the dice coefficient is less than the threshold value, the determination unit 108 determines that the related word candidate is not a related word, and excludes it from an object to be registered as a related word. In this manner, the determination unit 108 can exclude a general word (for example, “today”, “I”, etc.) from the registration target of the related word by performing threshold processing on a word having a small dice coefficient.

Dice coefficient sim = 2 × number of documents C / (number of documents A + number of documents B) (1)

  The related word registration unit 109 registers information including the place name and the related word candidate determined as the related word by the determination unit 108 in the related word storage unit 110.

  As illustrated in FIG. 5, the related word storage unit 110 stores information including a place name, a related word, and an area (area ID or area name). The information processing apparatus 100 outputs related words of place names (including landmarks) using information stored in the related word storage unit 110. Specifically, when the information processing apparatus 100 outputs a ranking of place names input to a blog or the like, the information processing apparatus 100 outputs related words of the place names. Thereby, the information processing apparatus 100 can output a related word of a place name corresponding to a popular place.

  An operation of the information processing apparatus 100 configured as described above will be described. FIG. 6 is a flowchart showing processing of the information processing apparatus 100. First, the document information acquisition unit 101 acquires document information, and the document information is stored in the document information storage unit 102 (S1). Then, the morphological analysis unit 103 performs morphological analysis on the document information document stored in the document information storage unit 102, and extracts words (S2). The area assigning unit 104 specifies a word including a place name in the document text from the processing result of the morphological analysis unit 103, and specifies an area corresponding to the place name. Then, information in which the area specified by the area assigning unit 104, the word of the document text extracted by each morphological analysis, and the document ID are associated is stored in the word storage unit 105 (S3). Further, the information processing apparatus 100 refers to information stored in the word storage unit 105, extracts the place name, and separately holds a reference place name group. Then, the information processing apparatus 100 refers to the information stored in the word storage unit 105 and separately holds a word group other than the place name as a related word candidate group.

  The candidate determination unit 106 uses any one place name in the reference place name group as the reference place name. Moreover, the candidate determination part 106 makes any one of a related word candidate group a related word candidate (S4). Then, the totaling unit 107 calculates the number of documents A (the number of documents including a place name n as a reference place name for a document set in a certain area p and a certain time width t) (S5), and the number of documents B (a certain area). p and the number of documents including the related word candidate word w are calculated for a document set in a certain time width t (S6), and the number of documents C (a document set in a certain area p and a certain time width t is targeted) The number of documents including place name n and word w) is calculated (S7). Then, the determination unit 108 calculates a dice coefficient using the document number A, the document number B, and the document number C (S8), and if the dice coefficient is equal to or greater than a predetermined threshold (S9; Yes), the related word Judge that the candidate is a related word. And the related word registration part 109 memorize | stores in the related word information storage part 110 as related word information the information with which the place name, the related word, and area ID were matched (S10). On the other hand, if the dice coefficient is less than the predetermined threshold (S9; No), the determination unit 108 determines that the related word candidate is not a related word and excludes it from the related word registration target (S11).

  Subsequently, the candidate determination unit 106 searches for a new related word candidate. Specifically, when not all related word candidates are determined for the reference place name (S12; Yes), the next related word candidate is determined as a determination target from the related word candidate group (S13), and step S5 is performed. Move to.

  When all the related word candidates are determined for the reference place name, that is, when there is no related word candidate that has not been determined for the reference place name (S12; No), the candidate determining unit 106 selects the next one from the reference place name group. Search the reference place name. Specifically, the candidate determination unit 106 determines whether or not there is a next reference place name in the reference place name group. If it is determined that there is a next reference place name (S14; Yes), the candidate determining unit 106 uses the place name as a reference place name, and selects any one related word candidate from the related word candidate group as a related word to be determined. It determines as a candidate (S15), moves to step S5, and determines whether or not the related word candidate is a related word based on the reference place name and related word candidate determined here.

  If there is no next reference place name in the reference place name group (S14; No), the candidate determining unit 106 determines that there is no word as a related word candidate, and ends the process.

  As described above, the information processing apparatus 100 calculates the number of documents A to C for all combinations of the reference place name and the related word candidate, and determines the related word corresponding to the place name using the number of documents. can do. That is, information for identifying a place name (unique name) included in a document of certain document information and determining whether the word is general as well as the number of documents including the word corresponding to the place name. Since it is determined whether the word is highly related to the place name, the related word can be appropriately extracted.

  In addition, since the information processing apparatus 100 targets only document information for a predetermined time width, for example, it is possible to specify a related word used most recently. That is, the information processing apparatus 100 can identify related words that are temporally characteristic.

  Furthermore, since the information processing apparatus 100 calculates the number of documents including related word candidates among the document information corresponding to the same area as the determination index information, it may be a word that is frequently used throughout the same area. For example, if it is frequently used in a limited part in a certain area, it is determined as a related word of the place name of the part. Therefore, the information processing apparatus 100 can extract related words that are words that are highly related to the place name, that is, related words that are geographically characteristic.

  Next, functions and effects of the information processing apparatus 100 and the related word determination method of this embodiment will be described. According to the information processing apparatus 100 of the present embodiment, the document information acquisition unit 101 acquires document information in which a document and time information are associated, and the morpheme analysis unit 103 analyzes the document of the acquired document information, The candidate determination unit 106 determines a related word candidate in the place name included in the document of the document information. The totaling unit 107 totals the number of documents including related word candidates and place names in the document information generated in a predetermined time span, and the determining unit 108 determines the result of the totaling and the related word candidates are general words. Whether or not the related word candidate is a related word of a place name is determined using determination index information that is index information for determining whether or not the related word is a related word. , Information relating the place name related word and the place name is registered. Thereby, the information processing apparatus 100 can extract an appropriate related word of the place name (unique name). That is, it is possible to remove a general word and extract a related word corresponding to a place name (unique name) in a predetermined time width.

  In the information processing apparatus 100 according to the present embodiment, the area assigning unit 104 identifies an area corresponding to the place name included in the document of the document information, and associates the document information with the area. In addition, the determination unit 108 uses, as determination index information, at least the number of documents including related word candidates in the document information in which the areas are associated in a predetermined time width. As described above, the information processing apparatus 100 determines whether or not the related word candidate is a related word of the place name based on the number of documents including the related word candidate in the document information associated with the area. The related words of the place name can be determined.

(Second Embodiment)
Next, an information processing apparatus 1A according to the second embodiment of the present invention will be described with reference to FIG. As shown in FIG. 1, only the totaling unit 107A and the determination unit 108A perform different processes among the components of the information processing apparatus 100 of the first embodiment, and other components are the same. Hereinafter, differences from the first embodiment will be mainly described.

  The counting unit 107A counts the number of documents including the place name (that is, the number of documents A) for a document set in an area to which the given place name belongs in a predetermined time width (for example, 3 days). The totaling unit 107A counts the number of documents including the word determined by the candidate determining unit 106 in a predetermined time width (hereinafter, the number of documents is also referred to as “document number B1”). Further, the totaling unit 107A calculates the number of documents including both the place name and the word determined by the candidate determining unit 106 (that is, the number of documents C) for a set of documents in the area to which the place name belongs in a predetermined time width. Count. Further, the totaling unit 107A counts the number of times the word determined by the candidate determining unit 106 appears in the document in the predetermined time width (hereinafter, the number of appearances is also referred to as “appearance number D”).

  The determination unit 108A determines whether the related word candidate is a related word of the place name. In the second embodiment, the dice coefficient is calculated using the document number A, the document number B1, and the document number C, which are the counting results counted by the counting unit 107A, and the document number B1 and the appearance count D are used. , Tf · idf is calculated, and a score obtained by calculating the dice coefficient and tf · idf is calculated (as a specific example, a score obtained by integrating the dice coefficient and tf · idf). If the calculation result is equal to or greater than the threshold value, the determination unit 108A determines the related word candidate as the related word of the place name, and the related word registration unit 109 relates the information that associates the place name, the related word candidate, and the area. Register in the word storage unit 110. On the other hand, if the calculation result of the dice coefficient is less than the threshold value, the determination unit 108 determines that the related word candidate is not a related word, and excludes it from an object to be registered as a related word.

  Subsequently, the operation of the information processing apparatus 100A according to the second embodiment will be described with reference to FIG. FIG. 7 is a flowchart showing the processing contents of the information processing apparatus 100A in the second embodiment.

  The processing contents of Steps S21 to S25, Step S27, and Steps S33 to S38 are the same as Steps S1 to S5, Step S7, and Steps S10 to S14 in the flowchart (FIG. 6) of the first embodiment, respectively.

As in the first embodiment, document information acquisition, morphological analysis, area assignment, and candidate determination are performed (S21 to S24). Then, the totaling unit 107 calculates the number of documents A (the number of documents including the place name n, which is the reference place name, for a document set in a certain area p and a certain time width t) (S25). Further, the totaling unit 107A calculates the number of documents B1 (the number of documents including the word w of the related word candidate for the document set in the time span t for all areas) (S26). Further, the totaling unit 107A calculates the number of documents C (the number of documents including the place name n and the word w for a document set in a certain area p and a certain time width t) (S27). Further, the tabulating unit 107A calculates the number of appearances D (the number of appearances of the word w as the related word candidate for the document set in the time span t for all areas) (S28). Then, the determination unit 108A calculates the dice coefficient based on the following equation (2) using the document number A, the document number B1, and the document number C (S29).

Dice coefficient sim1 = 2 × number of documents C / (number of documents A + number of documents B1) (2)

In step S30, the determination unit 108A calculates tf · idf using the number of documents B1 and the number of appearances D based on the following equation (3) (S30). Note that N in Equation (3) means the total number of documents.

tf · idf = number of appearances D × log (N / number of documents B1) (3)

  The determination unit 108A calculates a score from the dice coefficient sim1 and tf · idf (S31), and the determination unit 108A compares the score value with a threshold value to register the related word candidate as a related word. Is determined (S32).

  If the determination unit 108A determines that the score is equal to or higher than the threshold, the related word candidate is registered as a related word (S33), and if it is determined that the score is not equal to or higher than the threshold, it is not registered as a related word (S34). Then, the above process is repeated for all combinations of the reference place name and the related word candidate, and the related words for the place name are registered (S35 to S38).

  Next, operational effects of the information processing apparatus 100A according to the second embodiment will be described. In the information processing apparatus 100A of this embodiment, the determination unit 108 uses at least determination information as comparison information between the total number of documents (N) and the number of documents including related word candidates (document number B1) in a predetermined time width. Use. In this way, the information processing apparatus 100A refers to the comparison information between the total number of documents and the number of documents including related word candidates in a predetermined time width, thereby determining whether the related word candidate is a general word. Since it can be judged, the related word of a place name (unique name) can be judged appropriately.

(Third embodiment)
The information processing apparatus 100B according to the third embodiment of the present invention will be described with reference to FIG. FIG. 8 is a block diagram illustrating a functional configuration of the information processing apparatus 100B according to the third embodiment. Among the components of the information processing apparatus 100 according to the first embodiment, the counting unit 107B and the determination unit 108B are different, and further include a general word list generation unit 111.

  The general word list generation unit 111 lists words that appear frequently (for example, are included in a predetermined value or more) in the document of the document information acquired by the document information acquisition unit 101 as general words.

  The counting unit 107B counts the number of documents (document number C) including both the place name and a word that is a related word candidate for a set of documents in an area to which the given place name belongs in a predetermined time width.

  The determination unit 108B sorts for each place name in descending order of the number of tabulations by the tabulation unit 107B, generates a list of related word candidates whose total number is equal to or greater than a predetermined value, and among the generated list of related word candidates, Related word candidates not in the general word list are determined as related words.

  Next, the operation of the information processing apparatus 100B according to the second embodiment will be described with reference to FIG. FIG. 9 is a flowchart showing the processing contents of the information processing apparatus 100B in the third embodiment.

  The processing contents of steps S41 to S45 are the same as those of steps S1 to S5 in the flowchart (FIG. 6) of the first embodiment.

  Then, the totaling unit 107B calculates the number of documents C (the number of documents including the place name n and the word w for a document set in a certain area p and a certain time width t), and obtains a totaling result (S46).

  Subsequently, the candidate determination unit 106 searches for a new related word candidate. Specifically, when all the related word candidates are not determined for the reference place name (S47; Yes), the next related word candidate is determined as a determination target from the related word candidate group (S48), and step S46 is performed. Move to.

  When all the related word candidates are determined for the reference place name, that is, when there is no related word candidate that is not determined for the reference place name (S47; No), the candidate determining unit 106 searches for a new reference place name. Specifically, the candidate determination unit 106 determines whether there is a next reference place name in the reference place name group. If it is determined that there is a next reference place name (S49; Yes), the candidate determining unit 106 uses the place name as the reference place name and determines any one related word candidate from the related word candidate group (S50). Move to step S46.

  If there is no next reference place name in the reference place name group (S49; No), the candidate determination unit 106 has made a judgment on all sets of reference place names and related word candidates. Judge that there is no. The determination unit 108B sorts each place name in descending order of the number of tabulations by the tabulation unit 107B, generates a list of related word candidates whose total number is equal to or greater than a predetermined value (S51), and sets each related word candidate list. It is determined whether or not the word candidate is included in the general word list (S49). If it is not included in the general word list (S52; Yes), the related word registration unit 109 determines that the place name and the related word are included. Information related to the related word candidate and the area ID is stored in the related word information storage unit 110 as a related word (S53). On the other hand, if it is included in the general word list (S52; No), the determination unit 108B determines that the related word candidate is not a related word and excludes it from the target to be registered as a related word (S54).

  Next, operational effects of the information processing apparatus 100B according to the present embodiment will be described. In the information processing apparatus 100B of the present invention, the determination unit 108B uses general word list information generated based on the appearance degree in the document information as determination index information. In this way, the information processing apparatus 100B counts the number of documents including place names and related word candidates in a predetermined time width, and uses the general word list information generated based on the appearance degree of the document information in the document as determination index information. Since it is determined whether or not the related word candidate is a related word of the place name, it is possible to appropriately determine the related word of the place name.

  As mentioned above, although preferred embodiment of this invention was described, it cannot be overemphasized that this invention is not limited to the said embodiment.

  For example, in the above embodiment, the related word corresponding to the place name is registered. However, the present invention is not limited to this. For example, related words corresponding to other unique names (for example, personal names, movie titles, etc.) may be registered. It should be noted that the upper item of the person name is the person's classification information (affiliation organization, etc.), and the upper item of the movie is the genre of the movie.

DESCRIPTION OF SYMBOLS 100 ... Information processing apparatus, 101 ... Document information acquisition part, 102 ... Document information storage part, 103 ... Morphological analysis part, 104 ... Area addition part, 105 ... Word storage part, 106 ... Candidate determination part, 107 ... Aggregation part, 108 ... judgment part, 109 ... related word registration part, 110 ... related word storage part

Claims (5)

Document information acquisition means for acquiring document information associated with the document and time information;
Analyzing a document in the document information acquired by the document information acquisition means, and a related word candidate determination means for determining a related word candidate that is a related word candidate in the proper name included in the document;
Aggregating means for aggregating the number of documents including the related word candidate and the unique name among the document information generated in a predetermined time width;
The related word candidate is a related word of the unique name by using the counting result by the counting means and determination index information that is index information for determining whether the related word candidate is a general word. A determination means for determining whether or not,
An information processing apparatus comprising: registration means for registering information relating the related word of the unique name and the unique name if the result of the determination by the determining means is a related word. Further comprising an upper item specifying means for specifying an upper item of the unique name corresponding to the unique name included in the document in the document information, and linking the document information with the upper item;
The determination means uses, as at least the determination index information, the number of documents including the related word candidates in document information in which a higher-level item corresponding to the unique name is linked in a predetermined time width. The information processing apparatus according to claim 1.   The information processing apparatus according to claim 1, wherein the determination unit uses, as determination index information, comparison information between the total number of documents and the number of documents including the related word candidates in a predetermined time width.   The information processing apparatus according to claim 1, wherein the determination unit uses general word list information generated based on an appearance degree of the document information in the document as at least the determination index information. A document information acquisition step for acquiring document information associated with the document and time information;
Analyzing a document in the document information acquired in the document information step, and determining a related word candidate that is a related word candidate in a specific name included in the document; and
A totaling step of counting the number of documents including the related word candidate and the unique name in the document information generated in a predetermined time width;
The related word candidate is a related word of the unique name, using the counting result obtained by the counting step and determination index information that is index information for determining whether the related word candidate is a general word. A determination step for determining whether or not,
As a result of the determination in the determination step, if it is a related word, a registration step of registering information relating the related word of the unique name and the unique name;
A related word judgment method comprising:

Images