Please use this identifier to cite or link to this item:
Title: Automatically measuring subjectivity of Chinese sentences for sentiment analysis to reviews on the Internet
Other Titles: 面向互联网评论情感分析的中文主观性自动判别方法研究
Authors: Ye, Q
Zhang, Z
Law, R 
Keywords: Internet
Subjectivity detection
N-POS model
Issue Date: 2007
Publisher: 清华大学出版社
Source: 信息系统学报 (China Journal of Information Systems), 2007, v. 1, no. 1, p. 79-91 How to cite?
Journal: 信息系统学报 (China Journal of Information Systems) 
Abstract: As a new domain of unstructured data mining,online sentiment analysis has aroused great interest recently. Through automatically mining online reviews to certain products,consumers could know the distribution of attitudes of other consumers to this product before making a buying decision.Meanwhile,manufacturers and retailers would get consumers’feedback about their products or services,as well the opinion of customers to their competitors,which is useful for them to improve the products or services and gain competitive advantages.A crucial step before sentiment analysis is to identify subjective expressions in the context.Subjective sentences are usually the parts expressing sentiment,attitude or opinions in text.The existing researches on sentiment analysis mainly focus on English reviews. Few studies have been conducted to Chinese texts.As Chinese information has increased dramatically in cyber space, how to automatically analyze opinions of reviews in Chinese on the Internet has become urgent.One basic and important task is to establish a method to identify subjective expressions in Chinese reviews.This paper proposed an approach to measure subjective strength of Chinese sentences using patterns of continuous two words combination,2-POS model.In experiments of subjectivity classification to Chinese sentences,both subjective and objective sentences achieved high precision and recall.The results show that the performances of the proposed approach are comparable to existing studies in English.
Appears in Collections:Journal/Magazine Article

View full-text via PolyU eLinks SFX Query
Show full item record

Page view(s)

Last Week
Last month
Citations as of Dec 9, 2018

Google ScholarTM


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.