Discovering Playing Patterns: Time Series Clustering of Free-To-Play Game Data

摘要

The classification of time series data is a challenge common to alldata-driven fields. However, there is no agreement about which are the mostefficient techniques to group unlabeled time-ordered data. This is because asuccessful classification of time series patterns depends on the goal and thedomain of interest, i.e. it is application-dependent. In this article, we study free-to-play game data. In this domain, clusteringsimilar time series information is increasingly important due to the largeamount of data collected by current mobile and web applications. We evaluatewhich methods cluster accurately time series of mobile games, focusing onplayer behavior data. We identify and validate several aspects of theclustering: the similarity measures and the representation techniques to reducethe high dimensionality of time series. As a robustness test, we comparevarious temporal datasets of player activity from two free-to-play video-games. With these techniques we extract temporal patterns of player behaviorrelevant for the evaluation of game events and game-business diagnosis. Ourexperiments provide intuitive visualizations to validate the results of theclustering and to determine the optimal number of clusters. Additionally, weassess the common characteristics of the players belonging to the same group.This study allows us to improve the understanding of player dynamics and churnbehavior.