JP2010237867A - Metadata automatic imparting system and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a technology for automatically generating meta data from data described in an RDF(Resource Description Language) and OWL(Web Ontology Language) formats. <P>SOLUTION: A metadata generation part 11 analyzes an OWL(Web Ontology Language)/XML(eXtensible Markup language) document 50, and extracts the graph structure of an OWL/XML document 50 configured of subjects, predicates, and objects, and generates the metadata description of the OWL/XML document 50 based on the extracted graph structure, and generates an OWL/XML document 60 with metadata by adding the metadata description to the OWL/XML document 50. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a technique for generating an OWL / XML document with metadata by adding metadata to an OWL / XML document.

  In order to improve the convenience of information retrieval and utilization on the Internet, a concept of semantic WEB has been proposed in which each WEB page has meaning (metadata). The metadata description in the Semantic Web has standards such as RDF (Resource Description Language) and OWL (Web Ontology Language).

  Here, Patent Document 1 describes a technique for automatically converting schema information of structure description data into a schema structure of semantic description data.

Japanese Patent Laid-Open No. 2005-322025

  By the way, most of the current WEB sites do not have metadata. Therefore, it cannot cope with the semantic WEB. For this reason, in order to spread the semantic WEB, it is necessary to add metadata to the existing WEB site.

  However, it is necessary for an operator or the like to check the description of each WEB page XML (eXtensible Markup Language) or the like to perform the process of assigning the metadata accurately. This was an enormous amount of work, and this was one factor preventing the spread of semantic WEB.

  Therefore, an object of the present invention is to provide a technique for automatically generating metadata from data described in RDF and OWL formats.

  According to an embodiment of the present invention, there is provided an automatic metadata adding system that analyzes an OWL / XML document and extracts a graph structure of the OWL / XML document including a subject, a predicate, and an object, and the graph. Metadata generation means for generating a metadata description of the OWL / XML document based on the graph structure extracted by the extraction means, and OWL / XML with metadata in which the metadata description is added to the OWL / XML document Means for generating a document.

  In a preferred embodiment, the metadata generation means determines a literal type by applying a literal type inference rule to a literal subject and object out of the graph structure subject and object, and determines the literal type. A literal metadata description corresponding to the subject and object of the literal may be generated based on the type.

  In a preferred embodiment, the literal type inference rule is such that a regular expression is defined for each literal type, and the literal type is determined according to which regular expression the subject and object of the literal correspond to. Good.

  In a preferred embodiment, the metadata generation means generates a non-literal metadata description that declares a type of a non-literal subject and object node as a class among the subjects and objects of the graph structure. May be.

  In a preferred embodiment, the metadata description generation means applies a predicate metadata inference rule to the predicate of the graph structure to determine a predicate type, a domain, a range, and a multiplicity, and according to a result of the determination A predicate metadata description may be generated.

1 is a block diagram of an automatic metadata assignment system 1 according to an embodiment of the present invention. An example of a literal type inference rule and a predicate metadata inference rule is shown. An example of the OWL / XML document 50 is shown. It is a flowchart which shows the procedure of the metadata production | generation process which the metadata production | generation part 11 performs. The graph structure (RDF graph structure) 70 of the XML individual description 51 of the OWL / XML document 50 is shown. It is a flowchart which shows the detailed process sequence of a subject / object process. It is explanatory drawing of subject / object processing. It is a flowchart which shows the detailed process sequence of predicate processing. It is explanatory drawing of predicate processing. An example of the OWL / XML document 60 with metadata is shown.

  Hereinafter, an automatic metadata adding system for generating an OWL / XML document with metadata by adding metadata to an OWL / XML document according to an embodiment of the present invention will be described with reference to the drawings.

  FIG. 1 is a block diagram of an automatic metadata assignment system 1 according to this embodiment.

  The automatic metadata assignment system 1 is configured by, for example, a general-purpose computer system, and each component or function in the automatic metadata assignment system 1 described below is realized by executing a computer program, for example. . This computer program can be stored in a computer-readable recording medium.

  As shown in FIG. 1, the automatic metadata assignment system 1 reads an OWL / XML document 50 and generates a metadata-added OWL / XML document 60, a literal type inference rule storage unit 13, and the like. And a predicate metadata inference rule storage unit 15.

  The metadata generation unit 11 analyzes the OWL / XML document 50 and extracts a graph structure of the OWL / XML document 50 including a subject, a predicate, and an object. Further, the metadata generation unit 11 generates a metadata description of the OWL / XML document 50 based on the graph structure. For example, the metadata generation unit 11 performs meta-data processing based on the literal type inference rule stored in the literal type inference rule storage unit 13 and the predicate metadata inference rule stored in the predicate metadata inference rule storage unit 15. Generate a data description. Details of the metadata description generation will be described later. Further, the metadata generation unit 11 generates an OWL / XML document 60 with metadata in which a metadata description is added to the OWL / XML document 50.

  FIG. 2 shows an example of a literal type inference rule stored in the literal type inference rule storage unit 13 and a predicate metadata inference rule stored in the predicate metadata inference rule storage unit 15.

  FIG. 2A is an example of a literal type inference rule 130. As shown in the figure, in the literal type inference rule 130, a regular expression is defined for each literal type. For example, for types such as Boolean, integer,..., Regular expressions specific to each type are defined. The string type is difficult to abstract with regular expressions. Therefore, in this embodiment, a literal that does not correspond to any other type is a string type.

  FIG. 2B shows an example of the predicate metadata inference rule 150. As shown in the figure, the predicate metadata inference rule 150 defines an inference procedure for specifying a predicate type, a domain, a range, and a multiplicity.

  FIG. 3 shows an example of the OWL / XML document 50. The OWL / XML document 50 includes an XML individual description 51 as shown in FIG. As described in detail below, the metadata generation unit 11 analyzes the XML individual description 51 and generates metadata.

  FIG. 4 is a flowchart illustrating a procedure of metadata generation processing performed by the metadata generation unit 11. Hereinafter, detailed metadata generation processing will be described with reference to this flowchart.

  First, the metadata generation unit 11 receives an input of the OWL / XML document 50 (S100).

  Next, the metadata generation unit 11 analyzes the graph structure of the XML individual description 51 of the OWL / XML document 50 (S110). The graph structure analysis will be described with reference to FIG.

  FIG. 5 shows a graph structure (RDF graph structure) 70 of the XML individual description 51 of the OWL / XML document 50 shown in FIG.

  The XML individual description 51 can be described in a graph structure having three nodes “# 0”, “# 1”, and “# 2”, as shown in FIG. Each node has the following objects and predicates.

Node # 0 has the following three sets of objects and predicates.
(1) Object: “ABC Laboratory” 101 Predicate: “name” 201
(2) Object: Node # 1 Predicate: “hasGroup Corporations” 207
(3) Object: Node # 2 Predicate: “hasGroup Corporations” 209
Node # 1 has “SUB1” 103 as an object and “name” 203 as a predicate. The node # 2 uses “SUB2” 105 as an object and “name” 205 as a predicate.

  In the figure, the graph structure 70 is divided into a subject / object group 80 and a predicate group 90.

  Returning to FIG. 4, the metadata generation unit 11 performs subject / object processing based on the result of the graph structure analysis (S120). Detailed processing of subject / object processing will be described with reference to FIGS. FIG. 6 is a flowchart showing a detailed processing procedure of subject / object processing, and FIG. 7 is an explanatory diagram of subject / object processing.

  Since it has already been determined whether each node is a literal in the graph structure analysis of S110, first, in FIG. 6, the metadata generation unit 11 determines that the subject and object (subject / object group 80) of the graph structure 70 are literals. It is confirmed whether or not (S121).

  Then, the metadata generation unit 11 applies the literal type inference rule 130 to the literal subject and object of the subject and the object, that is, the subject and the object determined to be literals (S121: Yes). The literal type is determined (S123). At this time, the metadata generation unit 11 determines the literal type depending on which regular expression of the literal type inference rule 130 the subject and the object determined to be literals correspond to. And the metadata production | generation part 11 produces | generates the literal metadata description corresponding to the subject and object which were determined to be a literal based on the literal type determined by step S123 (S125).

  For example, in the example of FIG. 7, “ABC laboratory” 101 does not match any regular expression from Boolean to datetime of the literal type inference rule 130, so it is determined as the string type. The same applies to “SUB1” 103 and “SUB2” 105. As a result, the metadata generation unit 11 generates literal metadata descriptions 601, 602, and 603 as shown in FIG. For example, the metadata generation unit 11 may hold a template for literal metadata description in advance, and the template may be generated by applying a subject or object and a literal type to the template.

  Returning to FIG. 6, for the subject and object determined to be non-literal (S121: No), the metadata generation unit 11 declares the type of the node of the subject and object determined to be non-literal as a class. A literal metadata description is generated (S127).

  For example, in the example of FIG. 7, since the node # 0 is non-literal, the metadata generation unit 11 generates a non-literal metadata description 611 that declares the Corporation type that is the type of the node # 0 as a class. For example, the metadata generation unit 11 may hold a template of non-literal metadata description in advance and generate a node by applying a node type to the template.

  Returning to FIG. 4, the metadata generation unit 11 performs predicate processing based on the result of the graph structure analysis (S <b> 130). Detailed processing of the predicate processing will be described with reference to FIGS. FIG. 8 is a flowchart showing a detailed processing procedure of the predicate processing, and FIG. 9 is an explanatory diagram of the predicate processing.

  First, in FIG. 8, the metadata generation unit 11 applies a predicate metadata inference rule 150 to a predicate having a graph structure (predicate group 90), and determines the predicate type, domain, range, and multiplicity (S131). . For example, the metadata generation unit 11 determines the predicate type, domain, range, and multiplicity of each predicate according to the inference rules of the predicate metadata inference rule 150. And the metadata production | generation part 11 produces | generates the predicate metadata description according to the result of the determination (S133).

  For example, when the predicate metadata inference rule 150 is applied to the predicate “name” 201 as shown in the example of FIG. That is, the predicate type of the predicate “name” 201 is a data type because the object “ABC laboratory” 101 is a string type. Similarly, the domain of “name” 201 is a Corporation type, which is the type of the subject node # 0. The value range of “name” 201 is a string type which is a type of “ABC laboratory” 101 which is an object. The multiplicity is singular because the predicate name is “name”. Therefore, the metadata generation unit 11 generates the predicate metadata description 621 of the predicate “name” 201 based on the determination result. Also at this time, the metadata generation unit 11 may hold a template of the predicate metadata description in advance and generate the template by applying each determination result described above to the template.

  The predicate “hasGroup Corporations” 207 and 209 is also processed in the same manner as described above, and a predicate metadata description 622 common to these is generated.

  Returning to FIG. 4, the metadata generation unit 11 stores the literal metadata descriptions 601 to 603 and the non-literal metadata description 611 generated in step S120 and the metadata descriptions 621 and 622 generated in step S130. The XML document 50 is subscribed to generate an OWL / XML document 60 with metadata. An example of the OWL / XML document 60 with metadata is shown in FIG.

  As a result, an OWL / XML document 60 with metadata including metadata indicating the meaning of the OWL / XML document 50 can be automatically generated for an OWL / XML document 50 that does not include metadata. Therefore, metadata can be easily subscribed to an existing XML document, and an application using the metadata can be constructed. Therefore, it is possible to deal with semantic WEB and the like.

  The above-described embodiments of the present invention are examples for explaining the present invention, and are not intended to limit the scope of the present invention only to those embodiments. Those skilled in the art can implement the present invention in various other modes without departing from the gist of the present invention.

1 metadata automatic assignment system 11 metadata generation unit 13 literal type inference rule storage unit 15 predicate metadata inference rule storage unit 50 OWL / XML document 51 individual description 60 OWL / XML document document with metadata 70 graph structure 130 literal type inference Rule 150 Predicate Metadata Inference Rule

Claims (7)

A graph extracting means for analyzing an OWL / XML document and extracting a graph structure of the OWL / XML document comprising a subject, a predicate, and an object;
Metadata generation means for generating a metadata description of the OWL / XML document based on the graph structure extracted by the graph extraction means;
Means for generating a metadata-added OWL / XML document with the metadata description added to the OWL / XML document.   The metadata generation means applies a literal type inference rule to a literal subject and object out of the graph structure subject and object, determines a literal type, and determines the literal type based on the literal type. The automatic metadata assignment system according to claim 1, wherein literal metadata descriptions corresponding to the subject and object of the are generated.   The metadata according to claim 2, wherein the literal type inference rule has a regular expression defined for each literal type, and determines a literal type according to which regular expression a subject and an object of the literal correspond to. Automatic grant system.   2. The metadata according to claim 1, wherein the metadata generation unit generates a non-literal metadata description that declares a node type of a non-literal subject and object as a class among the subjects and objects of the graph structure. Automatic grant system.   The metadata description generating means applies a predicate metadata inference rule to the predicate of the graph structure to determine a predicate type, a domain, a range, and a multiplicity, and a predicate metadata description according to the determination result The metadata automatic assigning system according to claim 1, which is generated. A method of automatically assigning metadata performed by a computer,
Analyzing an OWL / XML document and extracting a graph structure of the OWL / XML document comprising a subject, a predicate, and an object;
Generating a metadata description of the OWL / XML document based on the extracted graph structure;
Generating an OWL / XML document with metadata to which the metadata description is added to the OWL / XML document. On the computer,
Analyzing an OWL / XML document and extracting a graph structure of the OWL / XML document comprising a subject, a predicate, and an object;
Generating a metadata description of the OWL / XML document based on the extracted graph structure;
Generating a metadata-added OWL / XML document with the metadata description added to the OWL / XML document; and a computer program for automatically assigning metadata.

Images