Detecting Relationship between Features and Sentiment Words using Hybrid of Typed Dependency Relations Layer and POS Tagging (TDR Layer POS Tags) Algorithm

Siti Rohaidah Ahmad, Mohd Ridzwan Yaakub, Azuraliza Abu Bakar


Through online product reviews, consumers share their opinions, criticisms and satisfactions on the products they have purchased. However, the abundance of product reviews may be confusing and time-consuming for prospective customers as they read and analyze differing views before buying a product. The unstructured format of product reviews needs a sentiment mining approach in analyzing customers’ comments on a product and its features. In this paper, the researchers explore and analyze the hybrid role of typed dependency relations (TDR) and part-of-speech tagging (POST) in detecting the relation between features and sentiment words. The researchers have also created a list of combination rules using TDR and POST to serve as a guide in identifying the relation between features and sentiment words in sentences. Results have shown that the hybrid algorithm could assist in identifying such a relationship and improve performance.


Typed dependency relations; part-of-speech tags; feature; sentiment word.

Full Text:




  • There are currently no refbacks.

Published by INSIGHT - Indonesian Society for Knowledge and Human Development