By Topic

Research on sentiment orientation of product reviews in Chinese based on cascaded CRFs models

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$33 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

3 Author(s)
Sheng-Chun Ding ; Sch. of Econ. & Manage., Nanjing Univ. of Sci. & Technol., Nanjing, China ; Ting Jiang ; Neng Wen

Computing sentiment strength of product reviews is a challenge in the field of sentiment analysis. The results obtained by standard condition random fields (CRFs) in the previous work were not well, so cascaded CRFs model is used to compute the strength of sentence sentiment in Chinese in this paper. The present approach contains two layers: in the first layer CRFs model is used to judge the polarity of the sentence, negative, positive or objective, then strength of polarity is obtained by CRFs model in the second layer. The proposed framework facilitates mapping the product reviews into five classes-(i.e. strong positive, general positive, objective, general negative and strong negative) by considering the label redundancy among layers. The authors choose context, conjunction, evaluated-word and the results of the polarity classification as the features of cascaded CRFs model. Experiments on the task3 corpus of COAE2008 show the remarkable performance of the proposed approach. Results of the first layer are added to the features in the second layers which can have the advantage to compute the sentiment strength of sentences.

Published in:

2012 International Conference on Machine Learning and Cybernetics  (Volume:5 )

Date of Conference:

15-17 July 2012