JP5250463B2 - Semantic association apparatus, processing method thereof, and program - Google Patents

Description

  The present invention provides a semantic associating device capable of performing processing for outputting designated comparison target information that is semantically matched with semantic determination target information constituting a database, without manual data creation work, and It relates to the processing method and program.

  In the case of outputting information that semantically matches the semantic determination target information among the information constituting the database in the computer system, the information processing apparatus that performs the processing is information that semantically matches the semantic determination target information. It is necessary to determine which information of the specified comparison target information. At this time, the information processing apparatus needs to know the meaning of each piece of information constituting the database in advance. For example, if the information represented by the same word as the word represented by the semantic determination target information does not exist in the specified comparison target information, the information processing apparatus designates the word by matching the character strings of the words. Information cannot be extracted from the compared information. Therefore, the semantic determination target information is matched with the specified comparison target information according to the meanings indicated by the respective words of the semantic determination target information and the specified comparison target information, and information having the same meaning is output from the database. There is a need.

  Here, Patent Document 1 is disclosed as a related technique. This Patent Document 1 extracts a correspondence relationship between a conversion source electronic document and a conversion destination electronic document, and as a result, a conversion procedure for converting information elements of the conversion source electronic document into information elements of the conversion destination electronic document. A technique for creating data and converting information elements of a conversion-source electronic document into information elements of a conversion-destination electronic document using the conversion procedure data is disclosed.

JP 2005-10976 A

  However, in the technique of the above-mentioned Patent Document 1, when creating the conversion procedure data for converting the information element of the conversion source electronic document into the information element of the conversion destination electronic document, the information element of the conversion source electronic document is changed. The element dictionary to be shown must be created manually, and this requires labor. Further, in the technique of the above-described Patent Document 1, when creating conversion procedure data for converting information elements of a conversion source electronic document into information elements of a conversion destination electronic document, Compare the element dictionaries of the information elements of the conversion destination electronic document, extract the correspondence between the conversion source electronic document and the conversion destination electronic document that match the keywords, and only convert the information elements between them Conversion procedure data is created. Accordingly, the information elements between the conversion source electronic document and the conversion destination electronic document are not automatically matched, and the keyword of each element dictionary of the information element of the conversion source electronic document and the conversion destination electronic document does not match. There is a possibility that it cannot be converted easily. In other words, the conversion source electronic document holds information elements called customer and the conversion destination electronic document holds user elements, and these information elements have almost the same meaning, but the keywords do not match. There is a problem that converting the information element of the document into the information element of the conversion-destination electronic document cannot be performed without manual work, and this also increases the labor of the conversion work.

  And, when outputting information that semantically matches the specified comparison target information among the information constituting the database, the procedure for manually creating data, such as the technique of Patent Document 1 described above, is omitted. It is desirable that the labor of the operator can be reduced.

  Therefore, the present invention has a meaning that the process of outputting the specified comparison target information that semantically matches the semantic determination target information among the information constituting the database can be performed without manual data creation work. It is an object of the present invention to provide an automatic association apparatus, a processing method therefor, and a program.

In order to achieve the above object, the present invention provides a first element name that represents an element of meaning determination target information with a word, and first definition information that defines the meaning of the word represented by the first element name. Meaning determination target information receiving means for receiving from the information provider the meaning determination target information stored in association with each other, a second element name represented by a word as a comparison target, and the second element name represented a second definition information represented by text to define the meaning of the compared information storage means for one or more stores comparative information held by at least association, in the second definition information the sentence and morphological analysis to extract a word representing the second definition information, wherein the first element information matching ratio calculation that calculates definition information words representing the a concordance rate of words representing the second definition information single indicating means, said first definition information Words representing the word match rate representing the second definition information and determines element information match rate determination means for determining more than the first threshold value, wherein the words representing the first definition information second definition information When the match rate is equal to or higher than the first threshold, the extracting means for extracting the set of the semantic determination target information and the comparison target information used for calculating the match rate, and the specified comparison target information and the set And an information output means for outputting the extracted meaning determination target information.

Further, the present invention provides a semantic determination target information receiving means for receiving from the information provider the semantic determination target information that holds the third definition information that defines the meaning of the word represented by the attribute information of the element of the semantic determination target information. Comparison object information storage means for storing one or a plurality of comparison object information holding fourth definition information represented by a sentence that defines the meaning of the word represented by the word attribute information as the comparison object ; A morphological analysis is performed on the sentence represented by the definition information to extract a word representing the fourth definition information, and a match rate between the word representing the third definition information and the word representing the fourth definition information is calculated. Attribute information match rate calculating means, attribute information match rate determining means for determining whether the match rate between the word representing the third definition information and the word representing the fourth definition information is greater than or equal to a second threshold, wherein the word representing the third definition information Word match rate indicating a fourth definition information, when the it is the second threshold or more, an extraction means for extracting a comparison target information and semantic determination target information used for calculation of their match rate, designated An information output unit that outputs the extracted semantic determination target information in combination with comparison target information.

According to the present invention, in the above semantic association device, the semantic determination target information is attribute information registered in the database of the information provider in association with the first element name, and the attribute information. 3rd definition information that defines the meaning of the word represented by, the comparison target information is attribute information registered in the database in association with the second element name, and the word represented by the attribute information And the fourth definition information represented by a sentence defining the meaning of the word, extracting the word representing the fourth definition information by performing morphological analysis on the sentence represented by the fourth definition information, based on the set of extracted semantic determination target information and the comparison information, and the attribute information match rate calculation means for calculating a matching rate of words representing the third and the fourth definition information words representing the definition information, single representing the third definition information The fourth word match rate indicating the definition information of and an attribute information match rate determination means for determining whether more than a second threshold value, the extraction means further words representing the third definition information And the meaning determination target information and the comparison target information used for calculating the match rate are extracted only when the match rate of the word representing the fourth definition information is equal to or higher than the second threshold value. .

Further, the present invention is the above semantic association device, wherein the extraction means has a match rate between the word representing the first definition information and the word representing the second definition information equal to or higher than the first threshold. Among the combinations of semantic determination target information and comparison target information, the combination of semantic determination target information and comparison target information having the highest match rate is extracted.

Further, the present invention provides the semantic association apparatus, wherein the extraction unit has a match rate between the word representing the third definition information and the word representing the fourth definition information equal to or greater than the second threshold. Among the combinations of semantic determination target information and comparison target information, the combination of semantic determination target information and comparison target information having the highest match rate is extracted.

According to the present invention, in the above semantic association device, at least the element name, the word representing the first definition information, and the word representing the third definition information are stored in the semantic determination target information. And a configuration determining means for determining whether or not there is.
Further, the present invention is an association result storage means for storing a combination of semantic judgment target information and comparison target information extracted by the extraction means, and an actual data change notification for attribute information of elements of the semantic judgment target information. When the change notification including the identification information of the attribute information of the element of the semantic determination target information is received, the semantic determination recorded in the association result storage unit based on the identification information included in the change notification Characterized by comprising post-evaluation processing means for identifying comparison target information to be combined with target information and rewriting the actual data about the attribute information of the comparison target information to the actual data received in the change notification. To do.

Further, the present invention is a processing method in the semantic association device, wherein the semantic determination target information receiving means of the semantic correlation device includes a first element name in which elements of the semantic determination target information are represented by words, The semantic determination target information that holds at least the first definition information that defines the meaning of the word represented by the first element name is received from the information provider, and the comparison target of the semantic correlation device The information storage means holds at least the second element name represented by a word as a comparison target and the second definition information represented by a sentence defining the meaning of the second element name in association with each other One or a plurality of pieces of comparison target information are stored, and the element information coincidence rate calculation means of the semantic association device performs morphological analysis on the sentence represented by the second definition information to obtain the second definition information. words representing extracted, the first Calculating a word representing the definition information concordance rate of words representing the second definition information, the semantic mapping element information match rate determination means of the device, said the word second representing the first definition information word word match rate indicating the definition information determines whether or first threshold, extraction means of the semantic mapping device, representing the word with the second definition information representing a first definition information When the match rate is equal to or higher than the first threshold, a set of semantic determination target information and comparison target information used to calculate the match rate is extracted, and the information output unit of the semantic association device The processing method is characterized in that the extracted semantic determination target information is output in combination with specified comparison target information.

The present invention is also a processing method in the semantic association device, wherein the semantic determination target information receiving means of the semantic correlation device defines the meaning of the word represented by the attribute information of the element of the semantic determination target information. 3 is received from the information provider, and the comparison target information storage unit of the semantic association device defines the meaning of the word represented by the attribute information of the word as the comparison target. One or a plurality of pieces of comparison target information holding the fourth definition information represented by sentences is stored, and the attribute information match rate calculation means of the semantic association device is represented by the fourth definition information. A morphological analysis is performed on the sentence to extract a word representing the fourth definition information , a matching rate between the word representing the third definition information and the word representing the fourth definition information is calculated, and the semantic association is performed. Device attribute information match rate A constant section, the third definition information representing words and matching rate of words representing the fourth definition information determines whether or second threshold value, the extraction means of the semantic mapping device, wherein the When the match rate between the word representing the definition information 3 and the word representing the fourth definition information is greater than or equal to the second threshold value, the semantic determination target information and the comparison target information used to calculate the match rate , And the information output means of the semantic association device outputs the extracted semantic determination target information in combination with the specified comparison target information.

According to the present invention, in the processing method described above, the semantic determination target information is attribute information registered in the information provider database in association with the first element name, and a word represented by the attribute information. Attribute information registered in the database in association with the second element name and the comparison target information of the semantic association device, and the attribute And the fourth definition information represented by a sentence defining the meaning of the word represented by the information, and the attribute information match rate calculating means of the semantic association device represents the sentence represented by the fourth definition information. Morphological analysis is performed to extract a word representing the fourth definition information, and based on a combination of the extracted meaning determination target information and comparison target information, the word representing the third definition information and the fourth definition information match of the words that describe the definition information Calculates, determines whether the attribute information match rate determination means of the semantic mapping device, the third and the word representing the definition information the fourth word match rate indicating the definition information of the second threshold value or more In addition, the extraction unit of the semantic association device further includes only when the match rate between the word representing the third definition information and the word representing the fourth definition information is equal to or greater than the second threshold. The semantic determination target information and the comparison target information used for calculating the match rate are extracted.

The present invention also holds at least the second element name represented by a word as a comparison target and the second definition information represented by a sentence defining the meaning of the second element name in association with each other. A computer of a semantic association device having comparison target information storage means for storing one or more comparison target information, a first element name representing elements of semantic determination target information in words, and the first element Meaning determination target information receiving means for receiving from the information provider the meaning determination target information that holds at least the first definition information that defines the meaning of the word represented by the name, and is represented by the second definition information. the sentences and morphological analysis to extract a word representing the second definition information, the first definition information element information matches to calculate the matching rate of words representing the the word second definition information representing Rate calculating means, the first constant Word and the word match rate representing the second definition information determines element information match rate determination means for determining more than the first threshold value representing the information, the second definition information words representing the first definition information Extraction means for extracting a set of semantic determination target information and comparison target information used for calculating the match rate when the match rate of the word representing is equal to or higher than the first threshold, specified comparison target information And a set of information output means for outputting the extracted meaning determination target information.

The present invention also provides comparison target information storage means for storing one or a plurality of comparison target information holding fourth definition information represented by sentences defining the meaning of the word represented by the word attribute information as the comparison target. Semantic determination in which the computer of the provided semantic association device receives the semantic determination target information holding the third definition information defining the meaning of the word represented by the attribute information of the element of the semantic determination target information from the information provider. object information receiving means, wherein the sentence represented in the fourth definition information by morphological analysis to extract a word representing the fourth definition information, the third said a word fourth representative of the definition information of the definition of attribute information matching ratio calculation means for calculating a matching rate of words representing the information, the third definition information words representing a concordance rate of words representing the fourth definition information attribute determines whether or second threshold value Information match rate determination means, Serial third definition information words representing a concordance rate of words representing the fourth definition information, when the it is the second threshold or more, compared with the mean determination target information used for calculation of their match rate A program that functions as an extraction unit that extracts information and an information output unit that outputs the extracted meaning determination target information in combination with specified comparison target information.

According to the present invention, the meaning determination target information defines attribute information associated with the first element name and registered in the database of the information providing source, and a meaning of a word represented by the attribute information. 3 definition information, and the comparison target information is represented by attribute information associated with the second element name and registered in the database, and a sentence defining the meaning of the word represented by the attribute information. a fourth definition information holding the, said computer further, the fourth by morphological analysis text represented by the definition information extracting word representing the fourth definition information, and the extracted meaning based on the combination of the comparison information and the judgment object information, the third attribute information match rate calculation means for calculating a word representing the definition information concordance rate of words representing the fourth definition information, the third It said a word that represents the definition information Attribute information match rate determination means for determining word matching rate if more than a second threshold representing a fourth definition information, to function as, the extraction means further wherein a word representing the third definition information first 4 is a program characterized by extracting meaning determination target information and comparison target information used for calculating the match rate only when the match rate of the word representing the definition information of 4 is equal to or higher than the second threshold. .

  According to the present invention, using the information stored in the provider information database 2 as it is, the association with the comparison target information registered in the comparison target information storage unit 13 corresponding to the provider information database 2 is performed. Is going. Therefore, the process of outputting the information stored in the provider information database 2 that matches the specified comparison target information semantically can be performed without manual data creation.

It is a block diagram which shows the structure of a semantic matching apparatus. It is a figure which shows the example of data of provider information. It is a figure which shows the example of data of comparison object information. It is a figure which shows the process outline | summary of the semantic matching apparatus 1. FIG. It is a 1st figure which shows the processing flow of a semantic matching apparatus. It is a 2nd figure which shows the processing flow of a semantic matching apparatus. It is a 3rd figure which shows the processing flow of a semantic matching apparatus.

Hereinafter, a semantic association apparatus according to an embodiment of the present invention will be described with reference to the drawings.
FIG. 1 is a block diagram showing the configuration of the semantic association apparatus according to the embodiment.
In this figure, reference numeral 1 denotes a semantic association device. Reference numeral 2 denotes a provider information database. Then, the semantic association device 1 includes a provider information receiving unit 11 (semantic determination target information receiving unit) that acquires the provider information from the provider information database, and a rule check that checks the data configuration rules of the provider information. The processing unit 12 (configuration determination unit), the comparison target information storage unit 13 (comparison target information storage unit) that stores the comparison target information, and the source information that has undergone the rule check and the comparison target information are compared and matched. An information comparison processing unit 14 (element information match rate calculation unit, attribute information match rate calculation unit) that performs processing such as rate calculation, and a comparison result evaluation unit 15 (element information match rate determination unit, element that performs comparison result evaluation processing) Attribute information match rate determination unit, extraction unit), post-evaluation processing unit 16 that performs post-processing on the provider information based on the evaluation result, and association evaluation information storage unit that stores a match rate threshold value 7, and a respective processing unit and the storage unit of.

FIG. 2 is a diagram illustrating an example of data of provider information.
Here, the provider information is information stored in the information provider database 2, as shown in FIG. 2, and includes an information element name (first information element name) and an information element keyword (first definition). Information), attribute names, attribute keywords (third definition information), and the like.
More specifically, if the information element name is information “rack”, this “rack” is a physical device that accommodates a physical device such as a computer or a communication relay device (router). Therefore, in the information providing source database 2, an information element keyword (first definition) that defines the meaning of the information element name such as “physical device, accommodation, physical device” is associated with the information element name “rack”. Information) is stored.
Further, if the information element name is information “rack” in the information provider database 2, the attributes of the “rack” include, for example, the height, length, depth, weight, type, and the like of the rack. The attribute name (height, length, depth, weight, type, etc.) relating to the rack, which is the information element name, is stored.
Furthermore, an attribute keyword that defines the meaning of the attribute name is registered in the information provider database 2 in association with the attribute name. For example, if the attribute name is “height”, the attribute keyword is information “rack, height” that defines the meaning of what the attribute name is (for “rack”) (“height”). It is.
It is assumed that information such as an information element keyword, an attribute name, and an attribute keyword is registered in the information provider database 2 for a plurality of various information element names.
In addition to the information on the information element name, information element keyword, attribute name, and attribute keyword described above, the provider information holds variable data corresponding to each attribute name. In other words, actual data (variables) about the attribute names “rack”, “height”, “length”, “depth”, “weight”, and “type” are stored. For example, actual data such as “height” = 2 m, “length” = 0.5 m, “depth” = 1 cm, “weight” = 20 kg, “type” = rack size, and the like.

FIG. 3 is a diagram illustrating an example of data of comparison target information.
The comparison target information is information for determining whether any information stored in the provider information database 2 is semantically matched as shown in FIG. Element name), information element definition (second definition information), attribute name, attribute definition (fourth definition information), and the like. These pieces of information are information corresponding to the information element name, information element keyword, attribute name, and attribute keyword of the provider information, respectively.
That is, if the information element name is information “rack”, this “rack” is a physical device that accommodates a physical device such as a computer or a communication relay device (router). The comparison target information storage unit 13 stores an information element definition that defines the meaning of the information element name such as “physical device, accommodation, physical device” in association with the information element name “rack”. Shall.
Further, if the information element name is information “rack” in the comparison target information storage unit 13, for example, the rack height, length, depth, weight, type, and the like can be used as the attributes of the “rack”. The attribute name (height, length, depth, weight, type, etc.) relating to the rack, which is the information element name, is stored.
Furthermore, an attribute definition that defines the meaning of the attribute name is registered in the comparison target information storage unit 13 in association with the attribute name. For example, if the attribute name is “height”, the attribute definition is information “rack, height” that defines the meaning of what the attribute name is (for “rack”) (“height”). It is.
It is assumed that information such as an information element definition, an attribute name, and an attribute definition is registered in the comparison target information storage unit 13 for one or a plurality of various information element names.

  The semantic association device 1 according to the present embodiment is provider information (semantic determination target information) acquired from an information provider (provider information database or the like), and information that represents elements of the information in words. Provided source information that holds at least an element name (first information element name) and an information element keyword (first definition information) that defines the meaning of the word represented by the information element name is received. Further, the semantic association apparatus 1 obtains an information element name (second element name) represented by a word as a comparison target and an information element definition (second definition information) that defines the meaning of the information element. One or a plurality of pieces of comparison target information stored in association with each other are stored for different information elements. Then, the semantic association device 1 calculates a match rate between the information element keyword indicated by the provider information and the information element definition indicated by the comparison target information, and the match rate is a threshold value of the definition information match rate (first threshold value). ) Determine whether it is above. Then, the semantic matching device 1 calculates the matching rate when the matching rate between the information element keyword indicated by the providing source information and the information element definition indicated by the comparison target information is equal to or higher than the threshold of the definition information matching rate. A set of the provider information and comparison target information used in the above is extracted. Then, the semantic association device 1 performs a process of outputting the specified comparison target information extracted in combination with the provider information as information semantically corresponding to the provider information.

  Further, the semantic association apparatus 1 according to the present embodiment includes the attribute information registered in the provider information database 2 (information provider database) in association with the provider information, the provider information, An attribute keyword (third definition information) that defines the meaning of the word represented by the attribute information is held, and the comparison target information is provided in association with the information element name (second element name) of the comparison target information It holds attribute information registered in the original information database 2 and the like, and an attribute definition (fourth definition information) that defines the meaning of the word represented by the attribute information. The semantic association device 1 is defined by the attribute keyword specified by the attribute information of the provider information and the attribute information of the comparison target information based on the combination of the extracted provider information and the comparison target information. The match rate of the attribute definition to be calculated is calculated, and it is determined whether the match rate is equal to or higher than the threshold of the attribute information match rate. Then, the semantic association device 1 has a case where the match rate of the attribute keyword specified by the attribute information of the provider information and the attribute definition defined by the attribute information of the comparison target information is equal to or greater than the threshold of the attribute information match rate Only, the provider information and the comparison target information used to calculate the match rate are extracted.

The process of outputting the specified comparison target information that semantically matches the provider information among the information stored in the provider information database 2 as information that corresponds semantically to the provider information by the above processing. It is an object of the present invention to provide a semantic association device 1 that can perform the above without manual data creation.
In addition, even when the character string of the information element name indicated by the provider information and the character string of the information element name indicated by the comparison target information do not completely match, the element keyword or attribute keyword corresponding to the information element name of the provider information Based on the information element definition and attribute definition for the information element name of the comparison target information, the semantic match rate of the information element name is calculated, and the specified comparison target information that semantically matches the provider information is automatically It is an object of the present invention to provide a semantic association device 1 that can be specified as follows.

FIG. 4 is a diagram showing an outline of processing of the semantic association apparatus 1.
As shown in this figure, the semantic association device 1 receives the provider information, acquires the provider information of a certain information element name from the provider information database 2, and stores it in the provider information. The process of extracting the existing information element name, information element keyword, attribute name, and attribute keyword is performed. The extracted information is compared with the information element name, information element definition, attribute name, and attribute definition stored in the comparison target information, and stored in the provider information database 2 based on the matching rate. Among the received information, the specified comparison target information that semantically matches the provider information is output as information that semantically corresponds to the provider information.

FIG. 5 is a first diagram illustrating a processing flow of the semantic association apparatus.
FIG. 6 is a second diagram showing the processing flow of the semantic association apparatus.
FIG. 7 is a third diagram illustrating the processing flow of the semantic association apparatus.
Next, details of the processing of the semantic association apparatus 1 will be described with reference to FIGS.
First, referring to FIG. 5, the semantic association device 1 receives an acquisition request for information stored in the provider information database 2 from, for example, an external device, and detects the start of processing based on this request. Then, the provider information receiving unit 11 of the semantic association device 1 outputs a provider information request signal to the provider information database 2 and receives certain first provider information. The provider information receiving unit 11 receives the provider information and transfers it to the rule check processing unit 12. Then, the rule check processing unit 12 performs a rule check process to determine whether the providing source information has a specified data structure.

  In this rule check process, first, the rule check processing unit 12 has a pair of description “information element definition START” and description “information element definition END” in the data of the provider information (FIG. 2). It is determined whether or not to perform (step S101). Further, the rule check processing unit 12 determines whether there is only one item defining the information element name in the data of the provider information (step S102). Further, the rule check processing unit 12 determines whether there is only one item defining the information element keyword in the data of the provider information (step S103). Further, the rule check processing unit 12 determines whether or not one or more items defining attribute names exist in the data of the provider information (step S104). Further, the rule check processing unit 12 determines whether one or more items defining attribute keywords exist in the data of the provider information (step S105). Further, the rule check processing unit 12 determines whether or not the number of items defining the attribute name and the number of items defining the attribute keyword are described in the provider information data (step S106). When the rule check processing unit 12 determines Yes in all the determinations of step S101 to step S106, the rule check processing unit 12 instructs the information comparison processing unit 14 to start processing and transfers the provider information. Further, when the rule check processing unit 12 determines No in any of the above-described steps S101 to S106, the rule check processing unit 12 outputs an error message or the like to the display unit or the like of the semantic association device 1 (step S107). ) And processing such as discarding the provider information to be determined.

  Next, when the information comparison processing unit 14 accepts the provider information, the information comparison processing unit 14 converts the information element name, information element keyword, attribute name, and attribute keyword stored in the data of the provider information in advance. Providers that are extracted based on the identification ID of provider information that has been added to identify information, and that the extracted information element name, information element keyword, attribute name, and attribute keyword are stored in a memory or the like Register in the information table 100 (step S201). Further, the information comparison processing unit 14 acquires the first comparison target information (FIG. 3) from the comparison target information storage unit 13 (step S202). In addition, the information comparison processing unit 14 performs morphological analysis of information element definition text stored in the comparison target information data, and extracts one or more words (step S203). Then, the information comparison processing unit 14 calculates a match rate between the word of the information element keyword extracted from the provider information and the word extracted from the information element definition described in the comparison target information data (step) S204). Here, as shown in FIG. 4, there are three information element keywords extracted from the provider information: “physical device”, “accommodation”, and “physical device”. It is assumed that “physical device”, “accommodation”, and “physical device” are included in the extracted words. In this case, (the number of words that match each word indicated by the information element keyword among the words extracted from the information element definition of the comparison target information) / (number of information element keywords) is calculated, and the matching rate is calculated. calculate. In the case of FIG. 4, among the words extracted from the information element definition of the comparison target information, the number of words that match each word indicated by the information element keyword is “3”, and the number of information element keywords is “3”. The match rate is 100%.

  Then, when the information comparison processing unit 14 finishes calculating the match rate between the word of the information element keyword of the provider information and the word extracted from the information element definition of the first comparison target information, the provider information The ID of the information element keyword extracted from the provider information, the identification ID of the first comparison target information, the word extracted from the information element definition of the comparison target information, and their match rates, Registration is made in the match rate totaling table 200 (step S205). Then, the information comparison processing unit 14 determines whether other comparison target information is recorded in the comparison target information storage unit 13 (step S206), and if it is recorded, the second comparison target information is displayed. The data is read from the comparison target information storage unit 13 (step S207). Then, the information comparison processing unit 14 calculates the match rate between the word of the information element keyword of the source information and the word extracted from the information element definition of the second comparison target information, and stores the match rate into the match rate totaling table 200. Register information in the same way. Similarly, the information comparison processing unit 14 sequentially reads all the comparison target information stored in the comparison target information storage unit 13 and compares it with the information element keyword word of the providing source information. The match rate of the word extracted from the information element definition of the target information is calculated, and the information is registered in the match rate totaling table 200.

  Then, the information comparison processing unit 14 sets an information element for each pair of the first providing source information received from the providing source information database 2 and all the comparison target information recorded in the comparison target information storage unit 13. When the match rate between the keyword word and the word extracted from the information element definition is calculated, the information comparison processing unit 14 notifies the comparison result evaluation unit 15 to perform the evaluation process. Then, the comparison result evaluating unit 15 calculates the highest match rate from the match rate totaling table 200 in which the calculated match rate exceeds the threshold of the element information match rate, and provides the highest match rate. A pair is extracted (step S208). In this embodiment, since the matching rate between the provider information and the comparison target information shown in FIGS. 2 and 3 is 100%, it is assumed that a pair of the provider information and the comparison target information is extracted. In addition, the comparison result evaluation unit 15 uses the provider information acquired from the provider information database 2 when none of the match rates in the match rate totaling table 200 exceed the threshold of the element information match rate. Error information indicating that there is no semantically matching comparison target information is output to the display unit, and the first provider information acquired from the provider information database 2 is discarded. If any of the calculated match rates exceeds the element information match rate, the information comparison processing unit 14 is notified of the identification ID of the pair of the extracted source information and comparison target information. Further, the comparison result processing unit 15 associates the information element name of the providing source information and the information element name of the comparison target information based on the pair of the providing source information and the comparison target information for which the highest match rate is calculated. And registered in the association result table 300 (step S209).

  From FIG. 7, the information comparison processing unit 14 then extracts the word indicated by the attribute name from the provider information table 100 based on the provider ID (FIG. 2) identification ID notified from the comparison result evaluation unit 15. The word indicated by the attribute keyword is read (step S301). The information comparison processing unit 14 extracts a word indicated by the attribute name of the comparison target information based on the identification ID of the comparison target information notified from the comparison result evaluation unit 15 (step S302). In addition, the comparison result processing unit 14 reads the text indicated by the attribute definition of the comparison target information based on the identification ID of the comparison target information notified from the comparison result evaluation unit 15 and performs morphological analysis on the document to perform attribute definition. Is extracted (step S303). Then, the information comparison processing unit 14 calculates a match rate between the attribute keyword word extracted from the provider information and the word extracted from the attribute definition described in the comparison target information data (step S304). .

  Here, as shown in FIG. 4, there are two attribute keywords 1 extracted from the provider information, “rack” and “height”, and among the words extracted from the attribute definition A of the comparison target information. Suppose that the words “rack” and “height” are included. In this case, (the number of words that match each word indicated by the attribute keyword 1 among the words extracted from the attribute definition A of the comparison target information) / (the number of attribute keywords 1) is calculated, and the match rate Is calculated. In the case of FIG. 4, among the words extracted from the attribute definition A of the comparison target information, the number of words that match each word indicated by the attribute keyword 1 extracted from the provider information is “2”. Since the number of extracted attribute keywords 1 is “2”, the match rate is 100%. As a result, the information comparison processing unit 14 writes information of the match rate 100% in the match rate totaling table 200 in association with the attribute name 1 of the provider information and the attribute name A of the comparison target information.

  Then, the information comparison processing unit 14 determines whether another pair of attribute name and attribute definition exists in the comparison target information (step S305), and if it exists, calculates a match rate for a word extracted from the attribute definition. The target is determined (step S306). For example, as shown in FIG. 4, in the comparison target information, a pair of attribute name B and attribute definition B corresponding thereto is described in addition to the attribute name and attribute definition pair of attribute definition A corresponding to attribute name A. ing. Therefore, the information comparison processing unit 14 calculates a match rate between the word of the attribute keyword 1 of the provider information and the word extracted from the attribute definition B of the comparison target information. As shown in FIG. 4, the attribute keyword 1 extracted from the provider information is “rack” and “height”, and “rack” is included in the words extracted from the attribute definition B of the comparison target information. ”And“ type ”are included. In this case, (the number of words that match each word indicated by the attribute keyword 1 among the words extracted from the attribute definition B of the comparison target information) / (the number of attribute keywords 1) is calculated, and the match rate Is calculated. In the case of FIG. 4, among the words extracted from the attribute definition B of the comparison target information, the number of words that match each word indicated by the attribute keyword 1 extracted from the provider information is “1”. Since the number of extracted attribute keywords 1 is “2”, the match rate is 50%. As a result, the information comparison processing unit 14 writes information on the match rate 50% in the match rate totaling table 200 in association with the attribute name 1 of the provider information and the attribute name B of the comparison target information.

  Then, the comparison result evaluation unit 15 specifies the attribute name A having the highest match rate as the attribute name in the data of the comparison target information corresponding to the first attribute name 1 described in the provider information, The attribute name 1 and the attribute name A are associated with each other and registered in the association result table 300 (step S307). That is, this specifies that the information corresponding to the attribute name 1 in the data of the provider information is the attribute name A in the data of the comparison target information.

  In addition, the information comparison processing unit 14 determines whether another pair of attribute name and attribute keyword exists in the provider information (step S308), and if it exists, the word extracted from the attribute keyword and the comparison target information The attribute keyword is determined as a match rate calculation target (step S309). That is, as shown in FIG. 4, in the provider information, a pair of attribute name 2 and attribute keyword 2 corresponding thereto is described in addition to the attribute name and attribute keyword pair of attribute keyword 1 corresponding to attribute name 1. ing. Therefore, the information comparison processing unit 14 calculates a match rate between the word of the attribute keyword 2 of the provider information and the word extracted from the attribute definition A of the comparison target information. As shown in FIG. 4, the attribute keyword 2 extracted from the provider information is “rack” and “type”, and “rack” is included in the word extracted from the attribute definition A of the comparison target information. , Suppose that the word “height” is included. In this case, (the number of words that match each word indicated by the attribute keyword 2 among the words extracted from the attribute definition A of the comparison target information) / (the number of attribute keywords 2) is calculated, and the match rate Is calculated. In the case of FIG. 4, among the words extracted from the attribute definition A of the comparison target information, the number of words that match each word indicated by the attribute keyword 2 extracted from the provider information is “1”. Since the number of extracted attribute keywords 2 is “2”, the match rate is 50%. As a result, the information comparison processing unit 14 writes information on the match rate 50% in the match rate totaling table 200 in association with the attribute name 2 of the providing source information and the attribute name A of the comparison target information.

  Further, as shown in FIG. 4, in the comparison target information, in addition to the attribute name / attribute definition pair of attribute definition A corresponding to attribute name A, a pair of attribute name B and corresponding attribute definition B is described. ing. Therefore, the information comparison processing unit 14 calculates a match rate between the word of the attribute keyword 2 of the provider information and the word extracted from the attribute definition B of the comparison target information. As shown in FIG. 4, there are two attribute keywords 2 extracted from the provider information, “rack” and “type”, and “rack” is included in the words extracted from the attribute definition B of the comparison target information. , “Type” word is included. In this case, (the number of words that match each word indicated by the attribute keyword 2 among the words extracted from the attribute definition B of the comparison target information) / (the number of attribute keywords 2) is calculated, and the match rate Is calculated. In the case of FIG. 4, among the words extracted from the attribute definition B of the comparison target information, the number of words that match each word indicated by the attribute keyword 2 extracted from the provider information is “2”. Since the number of extracted attribute keywords 2 is “2”, the match rate is 100%. As a result, the information comparison processing unit 14 writes information with a match rate of 100% in the match rate totaling table 200 in association with the attribute name 2 of the provider information and the attribute name B of the comparison target information.

  Then, the comparison result evaluation unit 15 identifies attribute names B having a high match rate as attribute names in the data of the comparison target information corresponding to the second attribute name 2 described in the provider information, and Attribute name 2 and attribute name B are associated with each other and registered in the association result table 300. That is, this specifies that the information corresponding to the attribute name 2 in the data of the provider information is the attribute name B in the data of the comparison target information.

  As a result of the above processing, the semantic association apparatus 1 stores an element name that matches the element name of the first provider information acquired from the provider information database 2 in the comparison target information storage unit 13. Further, in the pair of the provider information and the comparison target information, the attribute name described in each of the pair and the data contents of the attribute definition can be associated with each other. Then, the semantic association apparatus 1 requests the provider information database 2 to transmit the next provider information, and repeats the above-described processing using the second provider information. Thereby, it is possible to specify the comparison target information in the comparison target information storage unit 13 that semantically corresponds to all of the supply source information stored in the supply source information database 2.

  Then, the post-evaluation processing unit 16 uses the association result table 300 to identify the comparison target information corresponding to the provider information, and the actual data regarding the attribute name of the provider information is used as the attribute name of the comparison target information. For example, and is registered in the comparison target information storage unit 13 (step S310). That is, only information corresponding to the attribute name defined in the comparison target information in the provider information is reflected in the comparison target information.

  Then, the post-evaluation processing unit 16 performs subsequent processing using the association result table 300. That is, actual data corresponding to the attribute name is registered in association with the provider information in the provider information database 2. This value may be changed by processing of the provider information database 2. Then, the post-evaluation processing unit 16 receives, from the provider information database 2, for example, a notification of execution of the change process indicating the identification ID of the changed provider information and the attribute name holding the changed actual data. When the post-evaluation processing unit 16 receives the notification of the execution of the change process, the post-evaluation processing unit 16 acquires the changed real data information from the provider information stored in the provider information database 2. Further, the post-evaluation processing unit 16 compares the comparison target information registered in the association result table 300 in association with the identification ID of the provider information that holds the changed actual data, based on the notification of the execution of the change process. In addition, the attribute name on the comparison target information side that is associated with the attribute name and registered in the association result table 300 is identified from the attribute name corresponding to the changed actual data. Then, the actual data corresponding to the attribute name on the comparison target information side is rewritten with the changed actual data.

  When the information corresponding to the specified comparison target information is output from the provider information stored in the provider information database 2 by the processing of the semantic association device 1 as described above, Only the provider information that is semantically related to the specified information element name and attribute name is output.

  And according to the above-mentioned processing, even when the character string of the information element name indicated by the provider information of the provider database 2 and the character string of the information element name indicated by the comparison target information do not completely match, the provision is made Based on the element keyword and attribute keyword for the information element name of the original information and the information element definition and attribute definition for the information element name of the comparison target information, the semantic match rate of the information element name is calculated, and the source information and semantic It is possible to automatically specify the designated comparison target information that matches the above, thereby reducing the labor of the operator.

Note that, in the above-described processing of the semantic association device, the matching rate when the matching rate of the information element keyword indicated by the providing source information and the information element definition indicated by the comparison target information is equal to or greater than the threshold of the definition information matching rate. The pair of provider information and comparison target information used for calculation of the source information is extracted, and then the extracted result is defined by the attribute keyword specified by the attribute information of the provider information and the attribute information of the comparison target information The match rate of the attribute definition to be calculated is calculated, and when the attribute information match rate is equal to or higher than the threshold of the attribute information match rate, the provider information and the comparison target information used to calculate the match rate are extracted. However, in the process of the semantic association device, the former process may be omitted, and the provider information and the comparison target information may be extracted by only the latter process.
That is, the semantic determination target information is received from the information provider, the attribute keyword that defines the meaning of the word represented by the attribute information of the element of the semantic determination target information, and the meaning of the word represented by the attribute information of the word as the comparison target. When the match rate of the attribute definition to be defined is calculated, whether the match rate of the attribute keyword and the attribute definition is greater than or equal to a threshold (second threshold), and the match rate of the attribute keyword and the attribute definition is greater than or equal to the threshold Then, the semantic determination target information and the comparison target information used for calculating the match rate are extracted. Then, the semantic determination target information extracted in combination with the designated comparison target information may be output.

  In addition, the above-mentioned semantic matching apparatus has a computer system inside. Each process described above is stored in a computer-readable recording medium in the form of a program, and the above process is performed by the computer reading and executing the program. Here, the computer-readable recording medium means a magnetic disk, a magneto-optical disk, a CD-ROM, a DVD-ROM, a semiconductor memory, or the like. Alternatively, the computer program may be distributed to the computer via a communication line, and the computer that has received the distribution may execute the program.

  The program may be for realizing a part of the functions described above. Furthermore, what can implement | achieve the function mentioned above in combination with the program already recorded on the computer system, what is called a difference file (difference program) may be sufficient.

DESCRIPTION OF SYMBOLS 1 ... Semantic matching apparatus 2 ... Provider information database 11 ... Provider information receiving part 12 ... Rule check process part 13 ... Comparison object information storage part 14 ... Information comparison process Unit 15 ... Comparison result processing unit 16 ... Post-evaluation processing unit

Claims (13)

The meaning determination target that holds at least the first element name that represents the element of the meaning determination target information in words and the first definition information that defines the meaning of the word represented by the first element name Meaning determination target information receiving means for receiving information from an information provider;
One piece of comparison object information is stored in association with at least the second element name represented by the word as the comparison object and the second definition information represented by the sentence defining the meaning of the second element name. Comparison target information storage means for storing one or more;
Wherein said sentences are represented by the second definition information by morphological analysis to extract a word representing the second definition information, the word representing the said a word representing the first definition information second definition information Element information match rate calculation means for calculating the match rate;
Element information match rate determination means for determining whether the match rate between the word representing the first definition information and the word representing the second definition information is greater than or equal to a first threshold;
When the match rate between the word representing the first definition information and the word representing the second definition information is equal to or higher than the first threshold, the semantic determination target information and the comparison target used to calculate the match rate An extraction means for extracting a pair with information;
Information output means for outputting the extracted semantic determination target information in combination with specified comparison target information;
A semantic associating device comprising: Semantic determination target information receiving means for receiving, from an information provider, the semantic determination target information that holds the third definition information that defines the meaning of the word represented by the attribute information of the element of the semantic determination target information;
Comparison target information storage means for storing one or more pieces of comparison target information holding fourth definition information represented by sentences defining the meaning of the word represented by the word attribute information as the comparison target;
And morphological analysis of the sentence expressed by the fourth definition information extracting word representing the fourth definition information, the word representing the third and the fourth definition information words representing the definition information Attribute information match rate calculation means for calculating the match rate,
Attribute information match rate determination means for determining whether the match rate between the word representing the third definition information and the word representing the fourth definition information is equal to or greater than a second threshold;
When the match rate between the word representing the third definition information and the word representing the fourth definition information is greater than or equal to the second threshold value, the semantic determination target information and the comparison target used to calculate the match rate Extraction means for extracting information;
Information output means for outputting the extracted semantic determination target information in combination with specified comparison target information;
A semantic associating device comprising: The semantic determination target information is attribute information registered in the information provider database in association with the first element name, and third definition information that defines the meaning of the word represented by the attribute information. Hold
The comparison target information holds attribute information registered in the database in association with the second element name, and fourth definition information represented by a sentence that defines the meaning of the word represented by the attribute information. And
The sentence represented by the fourth definition information is subjected to morphological analysis to extract a word representing the fourth definition information. Based on the combination of the extracted meaning determination target information and comparison target information, the third and attribute information matching ratio calculation means for calculating a matching rate of words and word representing the definition information indicating the fourth definition information,
Attribute information match rate determination means for determining whether the match rate between the word representing the third definition information and the word representing the fourth definition information is equal to or greater than a second threshold;
The extraction means is further used to calculate the match rate only when the match rate between the word representing the third definition information and the word representing the fourth definition information is greater than or equal to the second threshold. The semantic matching device according to claim 1, wherein the semantic determination target information and the comparison target information are extracted. The extraction means includes a combination of meaning determination target information and comparison target information in which a match rate between the word representing the first definition information and the word representing the second definition information is equal to or higher than the first threshold. The semantic association device according to claim 1, wherein a combination of semantic determination target information and comparison target information having the highest match rate is extracted. The extraction means includes a combination of meaning determination target information and comparison target information in which a match rate between the word representing the third definition information and the word representing the fourth definition information is equal to or greater than the second threshold. The semantic association device according to claim 2 or 3, wherein a combination of semantic determination target information and comparison target information having the highest match rate is extracted. Configuration determination means for determining whether at least the element name, the word representing the first definition information, and the word representing the third definition information are stored in the meaning determination target information;
The semantic association device according to any one of claims 1 to 5, further comprising: Association result storage means for storing a combination of semantic determination target information and comparison target information extracted by the extraction means;
  Included in the change notification when the change notification of the actual data regarding the attribute information of the element of the semantic determination target information is received and includes the identification information of the attribute information of the element of the semantic determination target information Identifying comparison target information to be combined with the semantic determination target information recorded in the association result storage means based on the identification information to be obtained, and the actual data about the attribute information of the comparison target information in the change notification Post-evaluation processing means for rewriting the received actual data;
  The semantic association device according to any one of claims 2, 3, and 5. A processing method in a semantic association device,
A semantic determination target information receiving unit of the semantic association device defines a first element name representing an element of the semantic determination target information as a word and a meaning of the word represented by the first element name. The semantic determination target information that is stored in association with at least the definition information from the information provider,
The comparison object information storage means of the semantic association device has a second element name represented by a word as a comparison object and a second definition represented by a sentence defining the meaning of the second element name. Storing one or a plurality of pieces of comparison target information held in association with at least information;
The element information coincidence rate calculating means of the semantic association device extracts a word representing the second definition information by performing morphological analysis on the sentence represented by the second definition information, and the first definition calculating a matching ratio of a word representing the second definition information words representing the information,
The element information match rate determination means of the semantic association device determines whether the match rate between the word representing the first definition information and the word representing the second definition information is greater than or equal to a first threshold,
When the extraction means of the semantic association device has a match rate between the word representing the first definition information and the word representing the second definition information equal to or higher than the first threshold, Extract the set of semantic judgment target information and comparison target information used in the calculation,
The processing method characterized in that the information output means of the semantic association device outputs the extracted semantic determination target information in combination with specified comparison target information. A processing method in a semantic association device,
The semantic determination target information receiving means of the semantic association device receives the semantic determination target information holding third definition information defining the meaning of the word represented by the attribute information of the element of the semantic determination target information from the information provider. Receive,
The comparison object information storage means of the semantic association device has one or more pieces of comparison object information holding fourth definition information represented by sentences defining the meaning of the word represented by the attribute information of the word as the comparison object Remember multiple,
The attribute information match rate calculation means of the semantic association device extracts a word representing the fourth definition information by performing morphological analysis on the sentence represented by the fourth definition information, and the third definition calculating a matching ratio of the words representing the fourth definition information words representing the information,
The attribute information match rate determination means of the semantic association device determines whether the match rate between the word representing the third definition information and the word representing the fourth definition information is equal to or greater than a second threshold;
When the extraction means of the semantic association device has a match rate between the word representing the third definition information and the word representing the fourth definition information equal to or greater than the second threshold, Extract the semantic determination target information and comparison target information used in the calculation,
The processing method characterized in that the information output means of the semantic association device outputs the extracted semantic determination target information in combination with specified comparison target information. The semantic determination target information is attribute information registered in the information provider database in association with the first element name, and third definition information that defines the meaning of the word represented by the attribute information. Hold
The comparison target information of the semantic association device is attribute information registered in the database in association with the second element name, and a sentence that defines the meaning of the word represented by the attribute information. 4 definition information,
The attribute information coincidence rate calculating means of the semantic association device extracts a word representing the fourth definition information by performing morphological analysis on the sentence represented by the fourth definition information, and the extracted semantic determination target based on the set of information and the comparison information, and calculates the matching rate of the words representing the third and the fourth definition information words representing the definition information,
The attribute information match rate determination means of the semantic association device determines whether the match rate between the word representing the third definition information and the word representing the fourth definition information is equal to or greater than a second threshold;
The extraction means of the semantic association device further includes only when the match rate between the word representing the third definition information and the word representing the fourth definition information is equal to or greater than the second threshold. The processing method according to claim 8 , wherein the semantic determination target information and the comparison target information used for calculating the match rate are extracted. One piece of comparison object information is stored in association with at least the second element name represented by the word as the comparison object and the second definition information represented by the sentence defining the meaning of the second element name. A computer of a semantic association device provided with comparison target information storage means for storing one or more,
The meaning determination target that holds at least the first element name that represents the element of the meaning determination target information in words and the first definition information that defines the meaning of the word represented by the first element name Meaning determination target information receiving means for receiving information from an information provider,
Wherein said sentences are represented by the second definition information by morphological analysis to extract a word representing the second definition information, the word representing the said a word representing the first definition information second definition information Element information match rate calculation means for calculating the match rate,
Element information match rate determination means for determining whether the match rate between the word representing the first definition information and the word representing the second definition information is greater than or equal to a first threshold;
When the match rate between the word representing the first definition information and the word representing the second definition information is equal to or higher than the first threshold, the semantic determination target information and the comparison target used to calculate the match rate Extraction means for extracting a pair with information,
Information output means for outputting the extracted semantic determination target information in combination with the specified comparison target information;
A program characterized by functioning as Semantic correspondence provided with comparison object information storage means for storing one or more pieces of comparison object information holding the fourth definition information represented by a sentence defining the meaning of the word represented by the word attribute information as the comparison object Mounting device computer,
Meaning determination target information receiving means for receiving from the information provider the meaning determination target information that holds the third definition information that defines the meaning of the word represented by the attribute information of the element of the meaning determination target information;
And morphological analysis of the sentence expressed by the fourth definition information extracting word representing the fourth definition information, the word representing the third and the fourth definition information words representing the definition information Attribute information match rate calculation means for calculating the match rate,
Attribute information match rate determination means for determining whether the match rate between the word representing the third definition information and the word representing the fourth definition information is greater than or equal to a second threshold;
When the match rate between the word representing the third definition information and the word representing the fourth definition information is greater than or equal to the second threshold value, the semantic determination target information and the comparison target used to calculate the match rate Extraction means for extracting information,
Information output means for outputting the extracted semantic determination target information in combination with the specified comparison target information;
A program characterized by functioning as The semantic determination target information is attribute information registered in the information provider database in association with the first element name, and third definition information that defines the meaning of the word represented by the attribute information. Hold
The comparison target information holds attribute information registered in the database in association with the second element name, and fourth definition information represented by a sentence that defines the meaning of the word represented by the attribute information. And
Said computer further
The sentence represented by the fourth definition information is subjected to morphological analysis to extract a word representing the fourth definition information. Based on the combination of the extracted meaning determination target information and comparison target information, the third attribute information matching ratio calculation means for calculating a matching rate of words and word representing the definition information indicating the fourth definition information,
Attribute information match rate determination means for determining whether the match rate between the word representing the third definition information and the word representing the fourth definition information is greater than or equal to a second threshold;
Function as
The extraction means is used for calculating the match rate only when the match rate between the word representing the third definition information and the word representing the fourth definition information is greater than or equal to the second threshold. the program according to claim 11, which means the determination target information and extracting a comparison target information.

Images