1. Discovering Topics of Interest on Steam Community Using an LDA Approach
- Author
-
Yang Yu, Ba-Hung Nguyen, Fangyu Yu, and Van-Nam Huynh
- Subjects
Topic model ,Computer science ,Collaborative filtering ,Data science ,Domain (software engineering) - Abstract
Reviews from players regarding different characteristics of an esports game are one of the worthiest sources for the developers to enhance their services or adjust operating strategy. However, little research has been conducted on detecting esports players’ favorite topics dealing with topic modeling. Thus, this paper aims to use a data mining approach to analyze community data in the games domain available on Steam. We collected more than 1.2 million English reviews from four esports games up to August 2020 on Steam. Our contributions in this paper are: (i) we manually build a dataset by filtering out high-quality esports reviews, (ii) we then infer and group reviews into 3 groups with 19 topics, and (iii) we add more contributions to finding the emerging opinions of esports players towards the different topics of esports reviews, which might benefit further research on understanding esports reviews.
- Published
- 2021