Framework for automatically construct ontology knowledge base from semi-structured datasets

This paper propose the method to dynamically construct optimized ontology knowledge base (KB) using ontology schema model from semi-structured datasets based RDM (relational data model). In order to enhance a big data analysis ability, it must be provided classification or analysis method based on c...

Full description

Saved in:
Bibliographic Details
Published in2015 10th International Conference for Internet Technology and Secured Transactions (ICITST) pp. 152 - 157
Main Authors Gui-hyun Baek, Su-kyoung Kim, Ki-hong Ahn
Format Conference Proceeding
LanguageEnglish
Published Infonomics Society 01.12.2015
Subjects
knowledge base
Knowledge based systems
mapping description
Metadata
Ontologies
ontology
ProeprtyExp
Prototypes
Resource description framework
semi-structured dataset
Online AccessGet full text

Cover

More Information
Summary:This paper propose the method to dynamically construct optimized ontology knowledge base (KB) using ontology schema model from semi-structured datasets based RDM (relational data model). In order to enhance a big data analysis ability, it must be provided classification or analysis method based on context and meaning. Also recently, for providing high-level services on IoT environment, necessity of a semantic interaction between objects is emphasizing. An existing RDB is difficult to share or interact a various data. Besides, they are impediment in a contextual big data analysis or an interaction between individuals of IoT environment, because they has been operating by features of RDM that constructed a metadata or schema about domain. Therefore, a studies in constructing ontology KB that is possible contextual and semantic representation from existing RDB is in progress, as a result, typically, R2RML was published by W3C RDB2RDF Working Group. But, existing a mapping language can be limited, when RDB is not normalized (such as, 1NF), or when a semi-structured data provided (e.g. spreadsheet, text of key-value type (such as JSON), text of XML type etc.). Therefore, it has been requiring method that construct ontology KB from semi-structured datasets. We proposed a CVI (Cell-Value Importer) and a TTG (Transformation Table Generator) for dynamically importing value at semi-structured data. Also, we defined a Property Expression (PropertyExp) for describing mapping information by intuitive and concise sentence. Finally, we show that our design satisfied 10 of all 14 possible features of mapping languages [1] through testing by that constructed ontology KB from research equipment semi-structured dataset, after we implement prototype of this framework based JDK Environment.
DOI:10.1109/ICITST.2015.7412077