Learning a Model for Prediction of Game Leavers in Free-to-Play Games

Due to increasing popularity of video games and high number of products in the market, free-to-play games have become so popular. In free-to-play games, main source of incomefor developers is through advertisement and in-app purchase. Thus, one of the main concerns of developers it to increase and maintain the number of their active users in order to keeptheir income as high as possible. In this paper, an approach based on supervised learning isproposed that can be used for detecting the players that are about to leave the game. The main advantage of this approach is its generality and flexibility which makes it a suitableoption for different games. This research uses decision trees and artificial neural networksin the learning phase. This approach is used in the database of one of the most successful Iranian online free-to-play games called Sebghat . The results show the high performanceof this approach in modelling the behaviour of Iranian users in free-to-play games

کلیدواژه ها:

Data mining ، Game data mining ، Free-to-play games ، supervised learning