KR102121504B1 - System and method for building integration knowledge data base based a plurality of data sources - Google Patents

Abstract

A system for building a knowledge database according to the present invention includes: a service providing unit that provides service data requested to an application with reference to the knowledge database, a service data management unit that manages an ontology schema defined according to the service purpose; an interface unit that receives open data from different external data sources; and a data integration unit for generating a plurality of integrated data based on a data integration rule generated by analyzing the relationship between the plurality of open data and constructing the knowledge database based on the integrated data and the ontology schema. The data integration unit includes: an instance distributor for distributing a plurality of instances of the ontology schema for the knowledge database construction work based on curation module information representing the characteristics of a curation module and feedback information representing the result of the work performed by the curation module; a curation module group provided with a plurality of curation modules that individually perform the knowledge database construction work by receiving the plurality of instances distributed from the instance distributor, respectively; and a verifier for generating feedback information by verifying the result of the knowledge database construction work of the curation module group.

Description

System and method for building a knowledge database based on multiple data sources{SYSTEM AND METHOD FOR BUILDING INTEGRATION KNOWLEDGE DATA BASE BASED A PLURALITY OF DATA SOURCES}

The technical idea of the present invention relates to a knowledge database building system, and more particularly, to a knowledge database building system and method for building a knowledge database in an efficient manner based on data collected from a plurality of data sources. It is about.

The present invention is derived from research conducted by Saltlux Co., Ltd. as a part of the information and communication technology development project of the Ministry of Science and ICT. [Research period: 2018.01.01~2018.12.31, Research institute: Information and communication technology research promotion center, Research project name: WiseKB: Development of self-learning knowledge base and reasoning technology based on big data understanding, task identification number: 2013- 0-00109]

In the implementation of systems built by linked data research/commercial projects, data formats and protocols are being built in accordance with the standards established by the W3C. W3C proposed the recommended standard of'Linked Data Platform 1.0' from October 2012, and has been revised steadily until February 2015, and'Linked Data Platform 1.0' defines the definition format and communication protocol of Linked Data. . Based on these standards, various linked data conversion and management tools and services are being developed. In addition, development of a data service system that converts and manages open data stored in external databases into linked data has been actively conducted.

The problem to be solved by the technical idea of the present invention is to provide a knowledge database construction system and method for efficiently constructing a knowledge database based on a plurality of data sources.

The knowledge database construction system according to the present disclosure includes a service providing unit providing service data requested to an application with reference to a knowledge database, a service data management unit managing an ontology schema defined according to the service purpose, and external parties Based on the integrated data generated by analyzing the relationship between the interface unit receiving the open data from other data sources and the plurality of open data, and generating a plurality of integrated data, based on the integrated data and the ontology schema And a data integration unit for constructing the knowledge database, wherein the data integration unit is based on the curation module information indicating characteristics of the curation module and feedback information indicating a result of performing the operation of the curation module. For this, an instance distributor that distributes a plurality of instances of the ontology schema, and a plurality of curation modules that individually receive the plurality of instances distributed from the instance distributor and individually perform the knowledge database construction task are provided. And a validator for verifying the result of the knowledge database construction work of the curation module group and the curation module group, and generating the feedback information.

In addition, in one embodiment of the present disclosure, the instance divider divides the plurality of instances of the ontology schema according to the type of an instance that each of the curation modules can perform based on the curation module information, and the curation It is characterized by distributing to modules.

In addition, in one embodiment of the present disclosure, the instance distributor is characterized in that a part of the attributes corresponding to the instance of the ontology schema is selected in consideration of the type of the instance to distribute the construction task.

In addition, in one embodiment of the present disclosure, the curation module determines whether processing for a plurality of instances of the knowledge database construction job received from the instance distributor is possible, and the determination result is sent to the verifier. And providing the knowledge database based on the discrimination result.

In addition, in one embodiment of the present disclosure, the curation module selects an instance to add at least one processable instance in addition to the plurality of instances of the knowledge database construction job received based on the ontology schema Characterized in that it further comprises a wealth.

In addition, in one embodiment of the present disclosure, the curation module, when completing the task of constructing the knowledge database for the plurality of instances, provides a status output unit that provides a signal to the instance distributor to inform the completion of the task Further comprising, the instance distributor is characterized in that in response to the signal indicating the completion of the task, redistributing at least one instance that is not processed by another curation module in the curation module group to the curation module .

In addition, in one embodiment of the present disclosure, verification results are accumulated for each curation module in the curation module group, and information on performance of the plurality of curation modules is continuously updated based on the verification result, The feedback information is characterized by including information on the performance.

In addition, in one embodiment of the present disclosure, the verifier verifies the knowledge database construction work performed for each curation module in the curation module group, and provides the verification results to the curation modules, The curation modules are characterized by performing the knowledge database construction task again based on the verification result or completing the knowledge database construction task.

The knowledge database construction system according to an embodiment of the present disclosure provides efficient knowledge database construction by distributing instances of ontology schemas to the curation modules according to the characteristics of the curation module and the result of performing the operation of the curation module. There is an effect that can be performed.

1 is a block diagram schematically showing a knowledge database building system according to an embodiment of the present invention.
2 is a block diagram illustrating an interface unit according to an embodiment of the present disclosure.
3 is a block diagram illustrating a service data management unit according to an embodiment of the present disclosure.
4 is a diagram illustrating a data integration unit according to an embodiment of the present disclosure.
5 is a block diagram illustrating an instance distributor according to an embodiment of the present disclosure.
6 is a block diagram illustrating a group of curation modules according to an embodiment of the present disclosure.
7 is a block diagram illustrating a verifier according to an embodiment of the present disclosure.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. The embodiments of the present invention are provided to more fully describe the present invention to those skilled in the art. The present invention can be applied to various changes and may have various forms, and specific embodiments will be illustrated in the drawings and described in detail. However, this is not intended to limit the present invention to a specific disclosure form, and it should be understood that it includes all modifications, equivalents, and substitutes included in the spirit and scope of the present invention. In describing each drawing, similar reference numerals are used for similar components. In the accompanying drawings, the dimensions of the structures are enlarged or reduced than actual ones for clarity of the present invention.

The terms used in this application are only used to describe specific embodiments, and are not intended to limit the present invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this application, terms such as “include” or “have” are intended to indicate the presence of features, numbers, steps, actions, elements, parts or combinations thereof described in the specification, but one or more other features. It should be understood that the existence or addition possibilities of fields or numbers, steps, actions, components, parts or combinations thereof are not excluded in advance.

Further, terms such as first and second may be used to describe various components, but the components should not be limited by the terms. The terms may be used for the purpose of distinguishing one component from other components. For example, without departing from the scope of the present invention, the first component may be referred to as the second component, and similarly, the second component may also be referred to as the first component.

Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by a person skilled in the art to which the present invention pertains. Terms such as those defined in a commonly used dictionary should be interpreted as having meanings consistent with meanings in the context of related technologies, and should not be interpreted as ideal or excessively formal meanings unless explicitly defined in the present application. Does not.

Before describing embodiments of the present invention, terms and concepts to be used below will be briefly described.

Semantic technology is an intelligent technology that enables communication between computers by setting languages and rules that computers can understand, just like a person reads a screen and understands meaning. The semantic technology aims to express the relationship-semantic information (semanteme) between objects belonging to the environment in the form of an ontology that can be processed by a machine, that is, a computer. Ontology is a model that abstracts and shares what people think about things, and is a technique that is formalized and the types of concepts or constraints on use are explicitly defined. In the field of computer science, an ontology is a data model representing a specific domain, and is defined as structured data that describes concepts and relationships between concepts belonging to a specific domain. Ontology is a tool that can implement semantic technology, and is used as a tool to connect data semantically, and allows the computer to process and process the concept of objects in the form of a database.

Ontology components can be classified into instance, class, relationship, and property. We can explain that a class is a name we usually give to things or concepts. "Keyboard", "monitor", and "love" are all classes. On the other hand, an instance means itself in which objects or concepts appear in a practical form such as a concrete object or an event. In other words, "LG Electronics ST-500 Slim Keyboard", "Samsung Sync Master Wide LCD Monitor", "Romeo and Juliet's Love" can be considered as instances. Classes and instances can be very different depending on the application and purpose of use. That is, an object of the same expression can be a class in some cases and an instance in other cases.

Relationships refer to relationships that exist between classes/instances, and can be generally classified into a taxonomic relation and a non-taxonomic relation. The taxonomy relationship is a hierarchical expression that is classified into broader and concrete concepts for class/instance concept classification. For example, it is the "isA" relationship to indicate the inclusion relationship between concepts such as "people are animals." Relationships that are not classified are called unclassified. For example, "being healthy through exercise" is expressed using a "cause" relationship (causality).

Attributes are classes or instances that are associated with specific values to indicate a specific property/property of the class or instance. For example, to express "Samsung Syncmaster Wide LCD monitor is 24 inches.", an attribute such as hasSize can be defined.

In general, relationships and attributes are often not distinguished. Relationships/properties declared as relationships between actual classes/instances such as "isA (person, animal)", "cause (exercise, health)", "hasSize (Samsung Syncmaster Wide LCD, 24 inch)" Also called a property/property instance, this is for distinguishing the relationship/property defined and defined as "isA", "cause", and "hasSize".

In this field of semantic technology, the expression form triple is used as a means for expressing a relationship. Triple means expressing a concept in the form of a subject, a predicate, and an object. Each subject, predicate, and object can be expressed as a Uniform Resource Identifier (URI) of XML. RDF, OWL proposed by W3C, and TopicMaps proposed by ISO are standard languages that describe the semantic web ontology.

1 is a block diagram schematically showing a knowledge database building system 100 according to an embodiment of the present invention.

Referring to FIG. 1, the knowledge database construction system 100 includes an interface unit 110, a data integration unit 120, a service data management unit 130, and a service providing unit 140. The interface unit 110 may receive open data (OPD) from an external data pool. The interface unit 110 may provide a plurality of open data OPDs received from the data pool to the data integration unit 120 as an open data group OPDG. The data pool that provides open data to the interface 110 includes various data sources, such as the Internet, a database, cloud sourcing, and social networks. Can represent what can be created, retained, and distributed. In addition, the data pool may also include open data provided directly by the public or individuals. Open data may refer to public data disclosed for use by users. The type of open data (OPD) may be a static file such as CSV, Excel file, or RDF, or may correspond to a data catalog system type such as CKAN or an open API type.

Open data including informal data or unstructured data or formal data is received from a data pool. Unstructured data is data that is not implemented in a fixed form, and is contrasted with formal data or structured data including contents corresponding to a corresponding field. Although not stored in a fixed field, data including metadata or a schema, XML or HTML may be classified as semi-structured data, but the present invention uses semi-structured data as unstructured data. Note that it can be assumed as a work type.

The data integration unit 120 may receive an open data group (OPDG) including a plurality of open data, and convert the format of each open data into one unified format. In one embodiment, the open data may be in the form of data such as comma separated value (CSV), Excel, HypertextMarkup Language (HTML), Portable Document Format (PDF), Extensible markup language (XML). The data integration unit 120 may convert open data having various formats based on a conversion template to have a uniform format. The data integration unit 120 may analyze fields, formats, etc., which are structures of open data, and generate the conversion template based on the analyzed structures of open data. In one embodiment, the data integration unit 120 may uniformly convert the format of each open data received from the outside into an RDF (Resource Description Framework) data format.

The data integration unit 120 may analyze a relationship between each open data and generate integrated data rules. Based on the integrated data rule, integrated data incorporating open data may be generated. In one embodiment, open data may be converted into an RDF data format by the data integration unit 120, and integrated data incorporating open data converted into an RDF data format may have an RDF data format.

In order to provide a data providing service to a user, the service data management unit 130 may define an ontology schema that defines a relationship between space, time, user, service policy, and corresponding service according to the service purpose. The defined ontology schema may be stored in an ontology schema storage (not shown). For example, spatial ontology is a schema that can derive an administrative area from GPS coordinates, time ontology is a schema that can derive seasons, seasons, dawn, afternoon, etc. from input time information, healthcare service ontology is personal history, blood pressure , It can be defined as a schema that can grasp the health status from blood sugar measurement results. In addition, the service data management unit 130 may generate a mapping rule defining a method of mapping open data input from outside to an ontology schema. The service data management unit 130 may provide ontology information OI including at least one of an ontology schema and mapping rules to the data integration unit 120.

The data integration unit 120 according to an embodiment of the present disclosure may perform a knowledge database construction operation by applying integrated data to an ontology schema. The data integration unit 120 may include a plurality of curation modules capable of individually performing a knowledge database construction task for efficiency of a knowledge database construction task. In addition, the data integration unit 120 may include an instance distributor that distributes a plurality of instances of the ontology schema based on the curation module information indicating characteristics of the curation module and feedback information indicating the result of performing the operation of the curation module. have. Curation modules may perform knowledge database construction using instances distributed from the instance distributor. The data integration unit 120 includes a verifier for determining whether to apply knowledge data generated by the knowledge database construction task to the knowledge database by verifying the knowledge database construction task performed in the curation modules. Can be. The data integrator 120 may store the knowledge database created in a predetermined memory device and update the stored knowledge database.

In this way, the data integration unit 120 can perform an efficient knowledge database construction work by distributing instances of the ontology schema to the curation modules according to the characteristics of the curation module and the result of performing the work of the curation module. It works.

The data integration unit 120 may generate service data SD provided to the application 1 by referring to the knowledge database created based on the ontology schema and the integrated data defined by the service data management unit 130. In one embodiment, the data integration unit 120 may map integrated data to the ontology schema, generate mapping data, and generate service data SD from a knowledge database including the mapping data. In one embodiment, the mapping data may correspond to a triple having an RDF data format, and detailed description thereof will be described later. The service data SD may be stored in the service data management unit 130.

The service providing unit 140 receives a query for requesting service data from the application 1, analyzes the query, and then sends a service data request signal RS to the service data management unit 130 Can provide. In response, the service data management unit 130 provides the requested service data (RSD) to the service providing unit 140, and the service providing unit 140 provides the requested service data (RSD) to the application 1. Can be. Furthermore, so that various applications such as SPARQL Endpoint and RESTful API can access the data service system 100, the service providing unit 140 may include an interface corresponding to each of various applications.

The application 1 may provide service information necessary for defining a new ontology schema to the service data management unit 130 through the service providing unit 140, and the service data management unit 130 based on the new service information As a result, a new ontology schema can be defined.

2 is a block diagram illustrating an interface unit 110 according to an embodiment of the present disclosure. Referring to FIG. 2, in order to receive the first to nth open data (OPD_1 to OPD_n), the interface unit 110 uses the first to nth open data adapters 110_1 to 110_n corresponding to each data format. Includes.

The interface unit 110 may receive first to nth open data OPD_1 to OPD_n from the outside. The open data types of the first to nth open data OPD_1 to OPD_n may be different from each other. For example, the first open data OPD_1 may correspond to a static file format such as an Excel file, and the second open data OPD_2 may correspond to a data catalog system format such as CKAN, and n th open data ( OPD_n) may take the form of an open API. Accordingly, the interface unit 110 includes a first open data adapter 110_1 including a first interface corresponding to the static file format so that the first open data OPD_1 can be input, and the second open data The second open data adapter 110_2 including a second interface corresponding to the data catalog system format is included so that (OPD_2) can be input, and the n open data OPD_n is input so that the open is possible. An n-th open data adapter 110_n including an n-th interface corresponding to an API format may be included. However, the present invention is not limited thereto, and in order to receive open data having various open data formats, an open data adapter corresponding to the open data adapter may be further included. The interface unit 110 may provide a plurality of open data OPD_1 to OPD_n to the data integration unit 120 of FIG. 1 as an open data group OPDG at a time.

3 is a block diagram illustrating a service data management unit 130 according to an embodiment of the present disclosure. Referring to FIG. 3, the service data management unit 130 includes an ontology definition unit 131, a rule information generation unit 132, an ontology information storage unit 134, and a service data storage unit 135.

The ontology definition unit 131 may define at least one ontology schema defining spaces, times, users, service policies, and corresponding services, etc. according to the purpose of a service provided to an application. The ontology definition unit 131 may receive service information SI from an application, and define an ontology schema based on the service information.

The rule information generation unit 132 may generate a first mapping rule that defines a method of mapping the integrated data generated by integrating open data in the data integration unit 120 of FIG. 1 into an ontology schema. The first mapping rule may be generated based on the ontology schema defined in the ontology definition unit 131. In one embodiment, the first mapping rule is whether the triple resource of the integrated data having the RDF data format and the object of the defined ontology schema are the same group relationship, consent relationship, parent/child relationship, parent/child group relationship, It may indicate a specific relationship, such as a synonym relationship. However, this is not limited to this as an embodiment, and the specific relationship may indicate not only a consent relationship, but also a class relationship, a similar relationship, and the like. An object of an ontology schema means a class, an instance that is an object belonging to a class, and properties representing properties of the instance. Furthermore, when a data service system receives a query from an application, an inference rule for inference based on the query, that is, a user-defined rule for inference may be generated.

The ontology information storage unit 134 may store an ontology schema defined by the ontology definition unit 131. Also, the ontology information storage unit 134 may store the first mapping rule and the inference rule defined by the rule information generation unit 132. The ontology information storage unit 134 may provide the data integration unit 120 with the ontology information OI including at least one of the ontology schema and the first mapping rule to the data integration unit 120 of FIG. 1.

The service data storage unit 135 may receive and store the service data SD generated from the data integration unit 120 of FIG. 1. When the service data request signal RS is received from the service provider 140 of FIG. 1, in response, the requested service data RSD may be provided to the service provider 140. In one embodiment, the service data storage unit 135 may provide the service data SD as feedback information to the ontology definition unit 131 and the rule information generation unit 132. The ontology definition unit 131 may determine whether appropriate service data SD is generated based on the ontology schema by referring to the feedback information, and when the inappropriate service data SD is generated, the defined ontology schema may be modified. . In addition, the rule information generation unit 132 determines whether appropriate service data SD is generated based on the first mapping rule by referring to the feedback information, and when inappropriate service data SD is generated, it is defined. The first mapping rule can be modified.

4 is a diagram illustrating a data integration unit 120 according to an embodiment of the present disclosure.

Referring to FIG. 4, the data integration unit 120 may include a data format conversion unit 122, an RDF data storage unit 124, a knowledge database storage unit 126, and a data integration management unit 200.

The data integration unit 120 receives an open data group OPDG including a plurality of open data from the interface unit 110 of FIG. 1, and receives ontology information OI from the service data management unit 130 of FIG. 1. can do. The data integration unit 120 may build a knowledge database based on ontology information (OI) and integrated data, and generate service data (SD) from the knowledge database.

The data format converter 122 may include a structured data converter 122a and an RDF data converter 122b. The open data may be in the form of data such as comma separated value (CSV), Excel, HypertextMarkup Language (HTML), Portable Document Format (PDF), Extensible markup language (XML). The structured data conversion unit 122a may convert open data having various formats into a structured data type based on a conversion template. The structured data conversion unit 122a may analyze fields, formats, and the like of the structure of the open data (OPD), and generate the conversion template based on the structure of the analyzed open data (OPD).

The RDF data conversion unit 122b may convert open data converted into the structured data format to have an RDF data format. Furthermore, the RDF data converter 122b may directly receive an open data group (OPDG) and convert open data having different data formats to have an RDF data format. The open data converted to have an RDF data format from the RDF data conversion unit 122b may be stored in the RDF data storage unit 124.

The data integration management unit 200 may generate integrated data based on data integration rules generated by analyzing open data converted to have an RDF data format. For example, when integrating the first open data and the second open data converted to the RDF data format, the data integration rule agrees whether the triple resource of the first open data and the triple resource of the second open data are in the same group relationship. It may be a relationship, a parent/child relationship, a parent/child group relationship, or a specific relationship such as a synonym relationship. The specific relationship may indicate not only a consent relationship, but also a class relationship, a similar relationship, and the like.

The data integration management unit 200 may build a knowledge database based on the generated integrated data and ontology information (OI) (or ontology schema), and for this purpose, the instance distributor 210 and the curation module group ( 220) and a verifier 230. The curation module group 220 may include a plurality of curation modules that can individually perform a knowledge database construction task. Building a knowledge database may mean an operation of generating triple data that can be accessed with a specific identifier by connecting attributes, relationships with other instances, and classes for a plurality of instances included in the ontology schema.

The instance distributor 210 may distribute a plurality of instances of the ontology schema and provide them to respective curation modules in the curation module group 220. Specifically, the instance distributor 210 may distribute a plurality of instances of the ontology schema based on the curation module information indicating characteristics of the curation module. For example, the first curation module is specialized in performing knowledge database construction of instances related to the theme of'politics', and the second curation module is specialized in performing knowledge database construction of instances related to the theme of'celebrity'. When it is, the instance distributor 210 may distribute the instances in consideration of the characteristics of this curation module. The curation module group 220 may provide curation module information indicating characteristics of each curation module to the instance distributor 210, and the curation module information may be periodically updated.

In another embodiment, the instance distributor 210 may distribute a plurality of instances of the ontology schema based on feedback information generated from the verifier 230. The verifier 230 may perform verification using a verification module in which a certain reliability is guaranteed as to whether the knowledge database construction work of each curation module in the curation module group 220 has been well performed. The verifier 230 may provide feedback information including verification results for each curation module to the instance distributor 210. The verification result may include a result of whether each curation module performed the knowledge database construction work well and how quickly the knowledge database construction work was completed. For the future knowledge database construction work, the instance distributor 210 may adjust the number of instances distributed to each curation module based on the feedback information, and some curation modules that do not perform the knowledge database construction work well There may not be instances distributed.

In addition, the verifier 230 may control re-execution and suspension of the knowledge database construction work for each curation module in the curation module group 220 based on the verification result. The verifier 230 may allow the result to be applied to the knowledge database when the knowledge database construction work performed in the curation module group 220 is verified. The data integration management unit 200 may store the generated knowledge database in the knowledge database storage unit 126 or perform an update.

5 is a block diagram illustrating an instance distributor 210 according to an embodiment of the present disclosure.

Referring to FIG. 5, the instance divider 210 includes an edit instance selection unit 211, an edit attribute selection unit 212, a curation module designation unit 213, a curation module information storage unit 214, and feedback information collection It may include a portion 215. The edit instance selection unit 211 may select edit instances suitable for the ontology schema from the integrated data. The edit attribute selection unit 212 may select edit attributes suitable for each of the edit instances selected from the integrated data. For example, the editing attribute selector 212 may select, as an editing attribute, attributes (or relationships) related to the'regional district','policy','history', and the like most related to this when the selected editing instance is a member of the National Assembly. have.

The curation module designation unit 213 curates the selected editing instances based on the curation module information received from the curation module information storage unit 214 and the feedback information received from the feedback information collection unit 215, respectively. Can be distributed on. According to an embodiment, the curation module designation unit 213 may divide edit instances of the ontology schema according to a type of an instance each of the curation modules can perform and distribute the curation modules to the curation modules. In addition, the curation module designation unit 213 may adaptively distribute edit instances to each of the curation modules based on feedback information regarding the results of performing work of each curation module. For example, the curation module designation unit 213 may divide the selected edit instances into a first group instance and a second group instance, and provide the first curation module and the second curation module, respectively, and the first queue The migration module may perform a knowledge database construction task for the first group instance, and the second curation module may perform a knowledge database construction task for the second group instance.

The curation module information storage unit 214 may receive curation module information from the curation module group 220 of FIG. 4, and further, a history of edit instances distributed to each curation module ) Can be stored. The instance divider 210 may distribute edit instances based on the histories. In addition, the instance distributor 210 may receive information related to the current curation module status (eg, busy, idle) from each of the curation modules, and the status of the curation module Editing instances can be distributed based on information.

6 is a block diagram illustrating a curation module group 220 according to an embodiment of the present disclosure.

Referring to FIG. 6, the curation module group 220 may include first to Nth curation modules 220_1 to 220_N capable of performing individual knowledge database construction. Hereinafter, the configuration of the first curation module 220_1 will be mainly described, but it is clear that the configuration of the first curation module 220_1 can be applied to other curation modules 220_2 to 220_N.

The first curation module 220_1 may include an instance confirmation unit 211_1, an edit instance addition selection unit 212_1, a knowledge data construction unit 213_1, and a status output unit 214_1. The instance confirmation unit 211_1 may first check whether the edited instances are distributed or not. As described above, since each curation module 220_1 to 220_N may have a specialized instance theme, the instance confirmation unit 211_1 may determine whether the themes are editable editable instances based on the curation module information in which the theme appears. have.

The editing instance addition selector 212_1 checks whether there are more editing instances suitable for the ontology schema from the editing instances received from the instance distributor 210 of FIG. 5 based on the ontology schema, and adds at least one more editing instance. can do. The knowledge data construction unit 213_1 may perform a knowledge database construction operation based on distributed editing instances and ontology schema.

The status output unit 214_1 outputs the result of the knowledge database construction work to the verifier 230 of FIG. 4, and the instance distributor 210 generates status information indicating the current operation state of the first curation module 220_1 Can be output. The status output unit 214_1 may notify the instance distributor 210 of FIG. 5 when the construction of the knowledge database for the editing instances currently distributed to the first curation module 220_1 is completed. The instance distributor 210 of FIG. 5 responds to this notification, and there is additional editing instances (editing instances or knowledge database construction work confirmed to be impossible to be processed by other curation modules) in the first curation module 220_1. (Including failed edit instances).

7 is a block diagram illustrating a verifier 230 according to an embodiment of the present disclosure.

Referring to FIG. 7, the verifier 230 may include a non-construction confirmation unit 231, a knowledge data verification unit 232, a knowledge data application unit 233, and a feedback generation unit 234. The non-construction check unit 231 confirms this when the first curation module 220_1 cannot process the edited instances distributed by the instance confirmation unit 211_1 of FIG. 6, and checks the instance distributor 210 of FIG. 5 ) To redistribute the edited instances. In response to the request, the instance distributor 210 may redistribute the corresponding edit instances based on the status information of each curation module and the characteristics of the curation module.

The knowledge data verification unit 232 can perform verification by using the verification module with which reliability is guaranteed for the knowledge data (or knowledge database) generated by the curation modules 220_1 to 220_N of FIG. 6. have. The knowledge data verification unit 232 may control the knowledge data application unit 233 to apply the generated knowledge data to an existing stored knowledge database when the verification result is a pass.

The feedback generation unit 234 may generate feedback information including the verification result of the knowledge data verification unit 232 and provide it to the instance distributor 210 of FIG. 5. The feedback generation unit 234 counts the number of failures in the construction of the knowledge database of each curation module, and generates information indicating that the operation is impossible for the curation module exceeding the threshold value, to the instance distributor 210 of FIG. 5 Can provide. In response to this, the instance distributor 210 may not distribute the edited instances to the curation module corresponding to the non-workable object. The feedback generator 234 accumulates verification results for the curation modules 220_1 to 220_N of FIG. 6 to continuously update information about the performance of the curation modules 220_1 to 220_N based on the verification result. Can be. The performance information may include information such as processing speed for one instance and the number of failed construction tasks. The instance divider 210 of FIG. 5 may perform distribution for edit instances based on information about performance included in the feedback information.

Although the present invention has been described with reference to the embodiments shown in the drawings, these are merely exemplary, and those skilled in the art will understand that various modifications and equivalent other embodiments are possible therefrom. Therefore, the true technical protection scope of the present invention should be determined by the technical spirit of the appended claims.

Claims (8)

A service providing unit providing service data requested to the application with reference to the knowledge database;
A service data management unit that manages the service data based on an ontology schema defined according to a service purpose;
An interface unit that receives open data from different external data sources; And
And a data integration unit that generates a plurality of integrated data based on a data integration rule generated by analyzing the relationship of the plurality of open data, and builds the knowledge database based on the integrated data and the ontology schema,
The data integration unit,
An instance distributor for distributing a plurality of instances of the ontology schema for the knowledge database construction task based on the curation module information indicating characteristics of the curation module and feedback information indicating a result of performing the task of the curation module;
A curation module group provided with a plurality of curation modules that respectively receive the plurality of instances distributed from the instance distributor and individually perform the knowledge database construction task; And
And a verifier generating the feedback information by verifying the result of the knowledge database construction work of the curation module group. According to claim 1,
The instance distributor,
A knowledge database building system, characterized in that a plurality of instances of the ontology schema are divided and distributed to the curation modules according to the type of an instance that each of the curation modules can perform based on the curation module information. According to claim 1,
The instance distributor,
A knowledge database construction system characterized by distributing the construction task by selecting only a part of the plurality of attributes corresponding to the ontology schema instance in consideration of the instance type. According to claim 1,
The curation module,
It is determined whether processing of a plurality of instances of the knowledge database construction job received from the instance distributor is possible, and a determination result is provided to the verifier, and the knowledge database construction job is performed based on the determination result Knowledge database construction system, characterized in that. The method of claim 4,
The curation module,
The knowledge database construction system further comprises an instance adding selector for adding at least one processable instance in addition to the plurality of instances of the knowledge database construction job received based on the ontology schema. The method of claim 5,
The curation module,
When the knowledge database construction work for the plurality of instances is completed, further includes a status output unit that provides a signal to the instance distributor indicating that the work is completed,
The instance distributor,
Knowledge database construction system, characterized in that to redistribute the at least one instance that is determined to be non-processable from other curation modules in the curation module group in response to the signal indicating the completion of the operation. According to claim 1,
The verifier,
The verification results are accumulated for each curation module in the curation module group, and information about the performance of the plurality of curation modules is continuously updated based on the verification result,
The feedback information includes information about the performance, knowledge database construction system. According to claim 1,
The verifier,
Verify the knowledge database construction work performed for each curation module in the curation module group, and provide the verification results to the curation modules,
The curation modules,
Based on the verification result, the knowledge database construction system is performed again or the knowledge database construction system is completed.

Images