You can view the full text of this article for free using the link below.

Title: A supervised aspect level sentiment model to predict overall sentiment on tweeter documents

Authors: Syed Muzamil Basha; Dharmendra Singh Rajput

Addresses: School of Computer Science and Engineering, Vellore Institute of Technology, Vellore, 632014, India ' School of Information Technology and Engineering, Vellore Institute of Technology, Vellore, 632014, India

Abstract: As the social applications are gaining the more popularity, different kinds of social media platform are ready to publish and express the emotions in the form of reviews. In which, detecting the concealed sentiment patterns and discovering knowledge in the huge user-generated inputs is a challenging task and has greater social significance value. In traditional sentiment analysis approaches, statistical correlation between words is considered. Whereas, dependency factor between aspects and the sentiment words are ignored, which has a greater impact factor on overall sentiment analysis. In this paper, we propose a new supervised topic level sentiment model (SSM), which is capable of handling overall sentiment analysis problems. Belief maximisation algorithm is used in SSM model, and Dirichlet distribution is used to estimate aspects and sentiment words. In order to prepare and modernise new documents, a hyperparameter Gibbs sampling method is used. We conducted experiments on reviews related to different products in multiple documents and the results state that the SSM model outperforms the on-hand algorithm in terms of aspect recognition and overall sentiment prediction accuracy.

Keywords: belief maximisation; dirichlet distribution; Gibbs sampling.

DOI: 10.1504/IJMSO.2018.096451

International Journal of Metadata, Semantics and Ontologies, 2018 Vol.13 No.1, pp.33 - 41

Received: 01 Aug 2017
Accepted: 22 Aug 2018

Published online: 30 Nov 2018 *

Full-text access for editors Access for subscribers Free access Comment on this article