Sentiment Analysis of Product Reviews Using Bag of Words and Bag of Concepts

Abdulwahab Almestekawy ； Mustafa Abdulsalam

《International Journal of Electronics and Information Engineering》 11卷2期 (2019/12) Pp. 49-60

https://doi.org/10.6636/IJEIE.201912_11(2).01

摘要

Sentiment analysis is very useful for getting the overall attitude of people towards a specific topic. It can be thought of as an ability for the machine to have a common sense to judge people's opinions. This analysis becomes even more useful and efficient when the machine has access to large quantities of opinions and reviews towards the subject with which we are concerned. In this paper we focus on two quick methods of sentiment analysis and utilize them in the domain of product recommendation. The first method uses a naive Bayes classifier with bag of words approach. The second method uses a list of predefined keywords which we refer to as bag of concepts. Our results show that the proposed methods are time-efficient and can be used in online web services, while maintaining acceptable accuracy.

關鍵字

Bag of Concepts ； Bag of Words ； Sentic Lexicon ； Sentiment Analysis