WO2012057382A1

WO2012057382A1 - Ontology-based method for managing the integration of metadata

Info

Publication number: WO2012057382A1
Application number: PCT/KR2010/007556
Authority: WO
Inventors: 김평; 정한민; 이미경; 김선태; 강상규; 소대섭
Original assignee: 한국과학기술정보연구원
Priority date: 2010-10-29
Filing date: 2010-10-29
Publication date: 2012-05-03
Also published as: KR20120046884A; KR101183534B1

Abstract

Disclosed is an ontology-based method for managing the integration of metadata. The ontology-based method for managing the integration of metadata, in which different types of metadata are respectively integrated into standard metadata and managed, includes: obtaining each ontology for the standard metadata and for the metadata that will be integrated into the standard metadata; and converting the metadata into the standard metadata by mapping the information between items of the standard metadata and items of the metadata to be integrated into the standard metadata, and by applying conversion rules.

Description

Ontology-based Metadata Integration Management Method

The present invention relates to a metadata integrated management method, and more particularly, to an ontology-based metadata integrated management method.

The Korea Institute of Science and Technlogy Information (KISIT) generates and provides statistical information on research achievements (papers, patents, and research reports) in the field of science and technology. There are more than 30 sources of research achievements (publishers, domestic and foreign patents). Since bibliographic information and data format are different from each other, it is time-consuming and expensive to manage. In addition, metadata from the same source has been changing year by year, increasing management difficulties.

In order to provide an integrated search service for research achievements, it is necessary to convert each item of metadata by source into a standard metadata item.

In addition, for external distribution of research achievements and bibliography, it is necessary to convert and transmit standard metadata in various forms according to the format of the destination.

In the related art, a metadata integration system for integrating and managing different types of metadata has been studied.

One conventional method has been to apply a conversion rule to each of the metadata to be integrated into the standard metadata. In other words, when integrating arbitrary metadata into standard metadata, a conversion rule suitable for the metadata was created, converted, stored, and reused.

Because this method is unique to each of the metadata that the transformation rules want to integrate, it is impossible to apply a rule applied to one metadata to other metadata. Accordingly, there was a problem in that it could not be flexibly applied to the diversity and time-variability of metadata.

For example, even if the metadata A was converted to standard metadata but the format of the metadata A changed, the conversion rule had to be rewritten. On the other hand, it has a rule that converts metadata A to standard metadata, and even though metadata A and metadata B are structurally similar, metadata B is converted to standard metadata using metadata A's conversion rules. Couldn't be converted to, and had to have separate rules for converting metadata B to standard metadata.

An object of the present invention is to provide an ontology-based metadata integrated management method that can flexibly cope with diversity and time-varying metadata.

The present invention establishes an ontology for each of the metadata of the source and the metadata of the standard, maps the ontology, and applies the mapping information and conversion rules between the items of the metadata of the source and the items of the standard metadata. Allows to convert to standard metadata.

According to the metadata integrated management method according to the present invention, it is possible to flexibly cope with the diversity and time-variability of the metadata by applying the transformation information to the ontology rather than the metadata.

1 illustrates a preferred embodiment of the ontology-based metadata integrated management method according to the present invention.

FIG. 2 shows a detailed configuration of the data conversion step s106 shown in FIG.

3 shows an example of mapping data of a source with an ontology.

4 illustrates an example of mapping between ontology.

5 shows an example of a table to which a final conversion rule is applied.

Ontology-based metadata integrated management method according to the present invention to achieve the above object

In a method of integrating and managing metadata having different formats into standard metadata,

Obtain the ontology of each of the standard metadata and the metadata to be integrated therein, and convert the metadata into standard metadata by applying mapping information and conversion rules between the items of the standard metadata and the metadata of the metadata to be integrated therewith. Characterized in that.

Here, the metadata integrated management method according to the present invention

Generating an ontology of each of the standard metadata and the first metadata to be integrated into the standard metadata;

Mapping the ontology;

Generating a conversion rule for converting an item of the first metadata into an item of the standard metadata by referring to the mapping information; And

And converting the first metadata into metadata of the standard by using the ontology conversion rule.

The first metadata may be implemented in a database form, and an ontology corresponding to the first metadata may be created based on a table name, a field name, and a data type of the database type.

The first metadata may be implemented in the form of an Extensible Markup Language (XML) structured document, and an ontology corresponding to the first metadata may be configured in a data access path and attribute of the XML structured document. It is desirable to write based on this.

1 illustrates a preferred embodiment of the ontology-based metadata integrated management method according to the present invention. The method shown in FIG. 1 includes data acquisition step s102, data loading step s104, data conversion step s106, data approval step s108, management information addition step s110, indexing step s112 and data. It includes a distribution step (s114).

First, the data acquisition step (s102) is a process of obtaining metadata by source. In this case, the metadata includes one expressed in a database form and one represented in an XML structured document. These metadata are expressed in different formats and are the subject of integration into standard metadata.

In this case, the database type refers to a table created based on a table name, a field name, and a data type, such as a spreadsheet.

In addition, it is an XML structured document type that is created based on data access path (entity, attribute) and data type.

The data loading step s104 stores the database document and the XML structured document for each data source and analyzes the metadata item.

In the data conversion step s106, relationship information between the metadata item and the standard metadata item, generation of a conversion rule, and data conversion are performed. Specifically, the ontology of each metadata is obtained for each source, and the metadata of the source is converted into standard metadata by applying a conversion rule to the ontology.

The data approval step s108 is a step of checking the quality of the converted data. If the quality is not satisfactory or an error occurs in the data approval step s108, the data conversion step s106 is performed again to correct the error.

In step S110 of adding management information, a data item required for managing standard metadata is added.

Indexing step s112 is a step of adding an index for data retrieval and service provision.

The data distribution step s114 provides a data access service in response to an external data request based on the integrated metadata. In other words, it provides a search service for integrated metadata.

FIG. 2 shows a detailed configuration of the data conversion step s106 shown in FIG. The method shown in FIG. 2 discloses generating ontologies for metadata of each source and converting the metadata of each source into standard metadata through ontologies conversion.

The method illustrated in FIG. 2 includes an ontology generation step s202, an ontology relationship mapping step s204, a conversion rule generation step s206 with standard metadata, a conversion rule storage step s208, and a data conversion step s210. Include.

Ontology generation step (s202) generates the ontology for the source-specific data, for example bibliography of the article. Ontologies contain information that provides access to source-specific data. Database type metadata includes table name, field name, and data type. XML document type metadata includes data access path (entity, attribute) and data type.

The ontology relationship mapping step s204 generates relationship information between the ontology item of the new source and the ontology items of the existing ontology or standard metadata. Conventionally, although standard metadata and relationship information are unconditionally generated, in the present invention, a conversion rule with standard metadata may be generated through another ontology even if the standard metadata item is not used by using the characteristics of the ontology.

The generation of the conversion rule with the standard metadata step s206 generates a final conversion rule to the standard metadata ontology using the relationship information of the newly created ontology. In this case, the final conversion rule is generated by considering the relational properties of the ontology using the inference engine.

The conversion rule storage step s208 stores the final conversion rule in a separate database table. At this time, information about all procedures and conversion methods necessary for conversion should be included. Specifically, the conversion rule includes mapping information of the ontology, applied data conversion rule, and the like.

The data conversion step s210 converts source-specific metadata into standard metadata using the created conversion rule.

Here, the relationship attributes necessary for generating the conversion rule are as follows.

1) Ontology Basic Properties

RDF, RDFS and OWL properties

* rdfs: subPropertyOf, rdfs: subClassOf, owl: equivalentProperty, owl: equivalentClass

* rdfs: comment-adds information needed to understand the data

* rdf: seq-used to express order or combination when data consists of two or more items

2) Additional Definition Properties

-kisti: appliedRule

kisti: appliedRule defines the user function needed to transform the data.

For example, you define the functions needed to standardize the first and last name of the first name and how to apply them, or the functions required to standardize the format of the year. When two or more functions are applied, each function is marked sequentially.

kisti: originalPath

* kisti: originalPath describes data access information for accessing the original data. If the source type is a database, describe the table name and field name. If the source type is an XML document, describe the data access path.

-kisti: dataFormat

* kisti: dataFormat indicates the type of data. Stores the information needed to check the suitability of the data and convert the data from one form to another.

3 shows an example of mapping data of a source with an ontology.

The upper left side of FIG. 3 shows the source data 302, and the lower right side shows the ontology 304. Ontology 304 has classes of Journal 304a, Article 304b, Author 304c, Subject 304d, Publisher 304e, Institution 304f, Department 304g, Keyword 304h.

Each class is associated with an attribute table 306a-306h.

For example, Journal class 304a is associated with Journal attribute table 306a, and Journal attribute table 306a includes ID, printISSN, electronicISSN, journalTitle, journalAbbrTitle attributes.

Source data 302 is shown to include items related to Publish, articles related to Journal.

Referring to FIG. 3, it can be seen that each item of data 302 of origin is associated with a class and attribute table of ontology 304.

For example, journal class 304a is associated with Publisher class 304e and is also associated with Journal attribute table 306a as indicated by "hasPublisher."

When the source data 302 is in the form of a database, field entries in the database are considered as classes in constructing the ontology 304.

The mapping of data 302 of origin with ontology is described in detail as follows.

First, we associate metadata items by data source with attributes of ontology.

In this process, kisti: originalPath and kisti: dataFormat information of metadata are input.

kisti: origanalPath:

The data at the source is in the form of a database; Save table name and field name

The data at the source is in the form of an XML document; Store access (entity, attribute)

-kisti: dataFormat stores data format as well as structure information.

* In the case of the year, month, and date: YYYYMMDD, YYMMDD, etc., can be used to express and convert data values.

* For Name: Expresses notation such as last name, first name, etc.

4 illustrates an example of mapping between ontology.

Referring to FIG. 4, three

ontology

402, 404, 406 are shown. BL ontology 402 maps to Springer Ontology 404, and Springer Ontology 404 maps back to Standard Ontology 406. Here, BL Ontology 402 and Springer Ontology 404 are ontologies for the second and first metadata in the summary of the present invention, and Standard Ontology 406 is the ontology for standard metadata.

According to the present invention, as shown in FIG. 4, the object for ontology mapping may not necessarily be an ontology for standard metadata. That is, it can be directly mapped onto ontology for metadata of the existing standard, but can be freely mapped onto ontology for metadata of other sources.

For example, if Source A and Source B already have a relationship with the ontology for the metadata of the standard, that is, mapped, then the new source is an ontology item for the source A, source B, or the metadata of the standard. Can be related freely.

Referring to the lower left of FIG. 4, a case in which one item is divided into two or more items is illustrated. For example, the name attribute of the Author class of BL Ontology 402 is divided and mapped to givenName and fanilyName of the Author class of Springer Ontology 404.

As described above, when an item is divided into two or more items, it is processed as follows.

The name entry is separated by first and last name; Name-> intermediate node 408-> first name + last name is separated.

The intermediate node 408 specifies the function to be applied.

kisti: appliedRule: seperateName (String fullname)

Store the first and last names obtained by applying the function at the intermediate node 408 in the first name and last name entries, respectively. At this time, the order of storing in the intermediate node 408 is described in rdf: seq.

Meanwhile, referring to the right middle part of FIG. 4, a case in which two items are merged into one item is illustrated. For example, the volume_Start item and the volumeEnd item of the Article class of Springer Ontology 404 are combined and mapped to the volume item of the Article class of Standard Ontology 406.

-If year and month data are merged into year and month data, merge them into year + month date-> intermediate node-> year and month date.

The intermediate node 410 specifies the function to be applied.

-In case of integrating data without additional application function, it is not necessary to specify the application function.

The order in which the intermediate node 410 stores the data as a single item is indicated as an rdf: seq item.

5 shows an example of a table to which a final conversion rule is applied. The table of FIG. 5 has three feeds (Original Data, Standard Data, and Applied Rule).

Items of metadata of each source are written in the Original Data feed 502, items of standard metadata are written in the Standard Data feed 504, and items of standard metadata are written in the Applied Rules feed 506. Functions applied to convert an item into the form of an item of standard metadata are described.

For example, lower () converts to lowercase, and Substring (0,2) constructs a substring by extracting two characters from the first position (0) of the string.

Creating the final conversion rule is done as follows.

The conversion rule includes all the conversion information for converting the metadata into the final metadata item. If kisti: dataFormat information of data is different from each other, the function necessary to change dataFormat equally must be specified in kisti: appliedRule.

Where the ontology on which the source's metadata is represented is not directly linked to the ontology of the standard's metadata.

* Generate the final transformation rule by finding the ontology item for the standard metadata associated with the non-standard metadata item to which the metadata of the source is linked.

* If metadata A is associated with metadata B, and metadata B is associated with an item in the standard metadata ontology, the information required for metadata A to be converted to metadata B and metadata B to Using the necessary information, we create the final conversion rule for Metadata A to be converted to standard metadata.

If the ontology on which the metadata of the source is represented is directly linked to the standard metadata ontology, create the final conversion rule with the standard metadata item to which the metadata of the source is linked.

If metadata A is linked to metadata B, and metadata B is a standard metadata ontology, metadata A creates a final conversion rule containing the information needed to convert to metadata B.

As described above, according to the metadata integrated management method of the present invention, it is possible to efficiently manage metadata for each source by using an ontology relationship.

In addition, conversion rules can be more easily created by linking metadata items by source or existing metadata items.

In addition, the relationship between the ontology can be grasped at a glance through the ontology relationship.

In addition, when the metadata format of the existing source is changed, the new metadata is connected with the existing metadata, and only the changed metadata item is marked, so that the changed metadata item can be conveniently processed.

Claims

In a method of integrating and managing metadata having different formats into standard metadata,

Obtain the ontology of each of the standard metadata and the metadata to be integrated therein, and convert the metadata into standard metadata by applying mapping information and conversion rules between the items of the standard metadata and the metadata of the metadata to be integrated therewith. Metadata integration management method, characterized in that.
The method of claim 1,

Generating an ontology of each of the standard metadata and the first metadata to be integrated into the standard metadata;

Mapping the ontology;

Generating a conversion rule for converting an item of the first metadata into an item of the standard metadata by referring to mapping information of ontology; And

And converting the first metadata into metadata of the standard using the ontology conversion rule.
The method of claim 2,

The first metadata is implemented in a database form, and an ontology corresponding to the first metadata is created based on a table name, a field name, and a data type of the database type. .
The method of claim 2,

The first metadata is implemented in the form of an XML structured document, and the ontology corresponding to the first metadata is created based on a data access path (entity, attribute) and data type of the XML structured document. How to manage metadata integration.
The method of claim 2,

The conversion rule includes an ontology basic attribute, a function and application method for data type conversion, data access information for accessing original data, and data type.
The method of claim 2,

In the ontology generation process and ontology mapping process, when one item of metadata is divided into two items, the ontology is divided into two items through an intermediate node, and the function applied to the intermediate node and the order of storing in the intermediate node are converted. Metadata integration management method characterized by including in the rule
The method of claim 2,

In the ontology generation process and ontology mapping process, when two items are merged into one item, the ontology is integrated through the intermediate node, and the function to be applied to the intermediate node and the order of storing the single node in the intermediate node as a single item are included in the conversion rule. Metadata integration management method characterized in that