Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


Thesis_Minliang_Liao.pdf (1.62 MB)

ETD Abstract Container

Abstract Header

Analyzing and Predicting Helpfulness of Online Product Review

Liao, Minliang
http://orcid.org/0000-0001-6081-2770
http://rave.ohiolink.edu/etdc/view?acc_num=akron150059887755072

Abstract Details

2017, Master of Science in Engineering, University of Akron, Electrical Engineering.
Predicting the helpfulness of online product reviews is very important and useful in e-commerce; predicted helpfulness can be applied in recommendation systems and review rankings. Manual ranking of reviews requires huge human efforts. The goal of this thesis is to predict the helpfulness scores for online product reviews based on the information contained in the review text and to understand what makes a review helpful, and why. The scoring uses four semantic features, natural language processing, and machine learning techniques. Models are built to map reviews to their respective helpfulness scores, and results of reviews with different numbers of helpfulness votes are compared. In addition, the human annotation is done by having college students assign helpfulness scores to the selected reviews in the testing dataset; theses manually determined scores which are used for comparing against the reviews with automatic labels. The massive helpfulness votes on Amazon product reviews are used as ground truth. Experimental results show that two of the four semantic features can accurately help predict helpfulness scores and greatly enhance the performance compared with previously used features. Comparisons show the trained models align well with human perceptions and results on reviews with more votes are slightly better aligned. Semantic interpretation reveals that online customers prefer reviews that contain words that have positive/negative comments on the object described, that show reviewers are thoughtful and knowing when commenting on the object, and that emphasize emphasis on personal experience and positive emotions.
Bao Sheng, Dr (Advisor)
Sastry Shiva, Dr (Committee Member)
Tran Nghi, Dr (Committee Member)
58 p.
Electrical Engineering
helpfulness score; online reviews; natural language processing; machine learning; LIBSVM

Recommended Citations

Citations

  • Liao, M. (2017). Analyzing and Predicting Helpfulness of Online Product Review [Master's thesis, University of Akron]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=akron150059887755072

    APA Style (7th edition)

  • Liao, Minliang. Analyzing and Predicting Helpfulness of Online Product Review. 2017. University of Akron, Master's thesis. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=akron150059887755072.

    MLA Style (8th edition)

  • Liao, Minliang. "Analyzing and Predicting Helpfulness of Online Product Review." Master's thesis, University of Akron, 2017. http://rave.ohiolink.edu/etdc/view?acc_num=akron150059887755072

    Chicago Manual of Style (17th edition)

akron150059887755072
1,730
© 2017, all rights reserved.
This open access ETD is published by University of Akron and OhioLINK.