Clustering Player Behavioral Data and Improving Performance of Churn Prediction from Mobile Game
Hyoungjin Kwon, Wooyoung Jeong, Daewook Kim, Seong-Il Yang · 2018
Recognizing change of players' behavior is crucial as video game evolves. Therefore, clustering players' behavioral data has become an important issue. In this paper, we propose a clustering method from behavior log data. Using the proposed method, we present two experimental results: one is to analyze change of the clustered behavior pattern to in-game event and the other is to use the clustering result as a feature to improve a churn supervised learning model. Experimental results show that the proposed method has applied to analyze influence of in-game event on players' behavior as well as predict churn for real-world application.