Title: PROO ontology development for learning feature specific sentiment relationship rules on reviews categorisation: a semantic data mining approach

Authors: D. Teja Santosh; B. Vishnu Vardhan

Addresses: Computer Science and Engineering, JNTU Kakinada, Andhra Pradesh, India ' Computer Science and Engineering, JNTUH College of Engineering Jagtial, Telangana, India

Abstract: Crucial data like product features were obtained from consumer online reviews and sentiment words were gathered in Resource Description Format (RDF) in order to use them in meaningful reviews based categorisation on sentiments of the feature. The meaningful relationships among these pieces of RDF data are to be engineered in a Product Review Opinion Ontology (PROO). This serves as background knowledge to learn rule based sentiments expressed on product features. These semantic rules are learned on both taxonomical and non-taxonomical relations available in PROO Ontology. In order to verify the mined rules, Inductive Logic Programming (ILP) is applied on PROO. The learned ILP rules are found to be among the mined rules. The positively classified features are grouped to justify the goal of ILP for examples which are both complete and consistent. Left out negative examples are useful in knowing their count at the time of categorisation.

Keywords: product review opinion ontology; PROO ontology; semantic data mining; sentiment rules; inductive logic programming; ILP; review categorisation; product reviews; product features; consumer online reviews; RDF; rule-based sentiments.

DOI: 10.1504/IJMSO.2016.078105

International Journal of Metadata, Semantics and Ontologies, 2016 Vol.11 No.1, pp.29 - 38

Received: 23 Apr 2015
Accepted: 15 Jan 2016

Published online: 02 Aug 2016 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article