Automatic Ontology Construction using Conceptualization and Semantic Roles

Automatic Ontology Construction using Conceptualization and Semantic Roles

Amita Arora (YMCA University of Science and Technology, Department of Computer Engineering, Faridabad, India), Manjeet Singh (YMCA University of Science and Technology, Department of Computer Engineering, Faridabad, India) and Naresh Chauhan (YMCA University of Science and Technology, Department of Computer Engineering, Faridabad, India)
Copyright: © 2017 |Pages: 19
DOI: 10.4018/IJIRR.2017070104

Abstract

Ontologies are constructed to extract meaningful information from data sources. Constructing ontologies aim at capturing domain knowledge that gives a commonly agreed understanding of a domain, which may be reused, shared, among applications and groups. To ease the process of building ontologies automatically, this manuscript per the authors proposes a new approach which extracts semantic roles of nouns in the sentential structure along with usual concepts and their relationships. The extracted information about different roles, concepts and relationships among the concepts from different documents are then merged to construct an ontology for whole document. The proposed approach is implemented and the performance of the proposed technique is evaluated. Experiments show the ontology thus created captures most of the information given in the document.
Article Preview

Ontology building approaches can be classified according to the type of knowledge resources i.e. fully structured text such as databases, dictionaries or existing ontologies, semi-structured text such as HTML or XML or unstructured text such as plain text in journals, books, web etc.

Constructing ontology from fully structured text or semi-structured text involves updating the existing ontology or merging the existing ontologies to build a large ontology.

Harith Alani (2006) gives an approach for constructing ontologies from existing ontologies automatically using ontology mapping and merging techniques. Ontologies from a domain are ranked to get top ranked ontologies, segmented and are merged to form a detailed ontology. Semantic web technologies are used here for this purpose. Human intervention is required to choose the existing ontologies and to monitor when to stop merging the ontologies to have quality results.

Junli Li, Zongyi He and Qiaoli Zhu (2013) present a technique to merge geo-ontologies from various sources. The formal semantics are extracted using Formal Concept Analysis. Information entropy along with deviance analysis is used as a basis for reducing the size of merged concept lattice as preferred by the user. A threshold can be applied for reducing the merged concept lattice in accordance with the user interest. The difficulty with this method is that human intervention is largely expected to maintain the accuracy of merged ontology.

C.P. Abinaya and V.P. Sumathi (2013) utilize semantic and syntactic measures for merging and identifying similar concepts. WordNet is used here for determining similarities among classes and instances from different ontologies and then merging ontologies. The difficulty with this method is that merging can be done only for ontologies for a given domain.

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 9: 4 Issues (2019): Forthcoming, Available for Pre-Order
Volume 8: 4 Issues (2018)
Volume 7: 4 Issues (2017)
Volume 6: 4 Issues (2016)
Volume 5: 4 Issues (2015)
Volume 4: 4 Issues (2014)
Volume 3: 4 Issues (2013)
Volume 2: 4 Issues (2012)
Volume 1: 4 Issues (2011)
View Complete Journal Contents Listing