Hybrid Approach for User Reviews' Text Analysis and Visualization: A Case Study of Amazon User Reviews

International Journal of Interactive Mobile Technologies

article

Ruba Alnusyan¹  Ruba Almotairi²  Sarah Almufadhi²  Amal A. Al-Shargabi²  Jowharah F. Alshobaili² 

[1] Graduate student;Department of Information Technology, College of Computer, Qassim University

关键词: user reviews;    sentiment analysis;    topic modeling;    visualisation;   

DOI  :  10.3991/ijim.v16i08.30169

学科分类：社会科学、人文和艺术（综合）

来源: International Association of Online Engineering

【 摘 要 】

Nowadays, many people prefer to purchase through online websites. Usually, those people start with reading user reviews and comments before making a purchase decision. The user reviews are considered powerful sources of information about products, in which users share opinions and previous experiences on using these products. However, these reviews are mostly textual and uncategorized. Thus, new customers need to read a massive amount of reviews, one by one, to make a decision. This study attempts to bridge this gap and proposes a hybrid approach of topic modeling that combines supervised and unsupervised learning. In particular, the study collected a massive amount of Amazon user reviews, analyzed the reviews' texts, and combined two approaches of topic modeling, which are unsupervised and supervised learning, i.e., semi-supervised learning. Besides, the study makes classification on reviews based on sentiment analysis. The resulting reviews' topics and their sentiment classifications are displayed on a visual dashboard. The proposed hybrid approach showed better performance in terms of text analysis and clearer representation of review topics. The outcome of this study helps customers make their decision on purchase products in a more effortless and clearer way.

【 授权许可】

CC BY