Home // eKNOW 2014, The Sixth International Conference on Information, Process, and Knowledge Management // View article

Integrating Topic, Sentiment and Syntax for Modeling Online Review

Authors:
Rui Xie
Chunping Li
Qiang Ding
Li Li

Keywords: topic model; sentiment analysis; tag sentiment aspect model; online review analysis

Abstract:
The problem of analyzing online product reviews has drawn much interest of researchers. In this paper, we propose a novel probabilistic modeling framework based on Latent Dirichlet Allocation (LDA), which can reveal the latent aspect and sentiment of the review simultaneously. Unlike other topic models which only consider the text itself of online review, we firstly combine the Part-of-Speech (POS) tag into the model. We further propose three Tag Sentiment Aspect Models (TSA) to integrate the syntax information into modeling. The experiments show that our models are able to achieve a promising result not only on sentiment classification but on extraction of aspects of different sentiments.

Pages: 137 to 144

Copyright: Copyright (c) IARIA, 2014

Publication date: March 23, 2014

Published in: conference

ISSN: 2308-4375

ISBN: 978-1-61208-329-2

Location: Barcelona, Spain

Dates: from March 23, 2014 to March 27, 2014