|
Title:
|
CHURN PREDICTION FOR A MASSIVELY MULTIPLAYER ONLINE ROLE-PLAYING GAME USING A HIDDEN MARKOV MODEL |
|
Author(s):
|
Yongsuk Yang, Junseok Lim, Beom-suk Chung, Sungmin Lim, Sung-jun Lee, Yerim Choi, Jonghun Park, Jae Hoon Jin |
|
ISBN:
|
978-972-8939-77-9 |
|
Editors:
|
Piet Kommers, Tomayess Issa and Pedro Isaías |
|
Year:
|
2012 |
|
Edition:
|
Single |
|
Keywords:
|
Churn prediction, Hidden Markov model, Online games, Sequence mining, User classification |
|
Type:
|
Short Paper |
|
First Page:
|
265 |
|
Last Page:
|
269 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
Due to the increasing number of users playing massively multiplayer online role-playing games (MMORPGs), game companies need to keep the users playing their games because a large player-base is crucial for MMORPG games. There are many benefits if the game company can predict whether a user will stop playing the game or not, and if possible, the reason why the user is leaving. In this paper, we propose a model which can predict whether or not a user will exit the game by using a sequence mining method and a hidden Markov model (HMM). We also utilize the HMM to analyze the reason behind the users churn. Experiment results show that the model effectively classifies the exit users from the staying users. |
|
|
|
|
|
|
