Natural language processing driven ontology modelling to address semantic interoperability in information extraction Online publication date: Wed, 12-Sep-2018
by B. Damayanthi Jesudas; B. Gurumoorthy
International Journal of Product Lifecycle Management (IJPLM), Vol. 11, No. 3, 2018
Abstract: The aim of this research is to investigate the use of natural language processing techniques for ontology modelling that would address semantic interoperability problems arising due to terminological and semantic differences during information extraction from documents. These documents contain terminologies that are specific to the group and experts involved in creating them. As manual extraction is a tedious process, natural language processing techniques are suggested in this work. Text from different documents describing new terms are taken as input. These texts are processed to establish either semantic equivalence of the new term with concepts in reference ontology or add the term as a new concept if appropriate. Semantic role labelling is used for concept modelling and semantic role labelling tool is used in deriving the concept instances from the text, which are matched against concepts in the reference ontology to establish equivalence. If instances of attributes are equivalent, then concepts are said to be equivalent and if equivalence is limited to attributes, then ontology is populated with new concept. Results on technical text are presented with discussion on how the limitations observed can be removed.
Online publication date: Wed, 12-Sep-2018
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Product Lifecycle Management (IJPLM):
Login with your Inderscience username and password:
Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.
If you still need assistance, please email email@example.com