Behavlet Analytics for Player Profiling and Churn Prediction

Darryl Charles, Benjamin Ultan Cowley

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Players exhibit varying behaviour from each other when playing games. Indeed, a player’s own behaviour will change as they learn to play and changing behaviour may also be indicative of when they are prematurely about to quit playing permanently (known as ‘churn’). There can be many reasons for player churn including finding a game too easy, too hard, or just not understanding what they must do. The accurate prediction of player churn is important as it allows a publisher or developer to understand and intervene to improve retention and thus increase revenue. Profiling player behaviours through their actions in-game can facilitate personalization, by adapting gameplay for different types of player to enhance player enjoyment and reduce churn rate. Behavlets are data-features that encode short activity sequences (‘atoms’ of play), which represent an aspect of playing style or player personality traits, e.g. aggressiveness or cautiousness tendencies. Previously we have shown how Behavlets can be used to model variation between players. In this paper, we focus on Behavlet sequences and how process mining and entropy-based analysis can profile evolving behaviour, predict player churn, and adapt play to potentially increase enjoyment.

Original languageEnglish
Title of host publicationHCI International 2020 – Late Breaking Papers
Subtitle of host publicationCognition, Learning and Games - 22nd HCI International Conference, HCII 2020, Proceedings
EditorsConstantine Stephanidis, Don Harris, Wen-Chin Li, Dylan D. Schmorrow, Cali M. Fidopiastis, Panayiotis Zaphiris, Andri Ioannou, Andri Ioannou, Xiaowen Fang, Robert A. Sottilare, Jessica Schwarz
Pages631-643
Number of pages13
Volume12425
DOIs
Publication statusPublished - 4 Oct 2020
Event22nd International Conference on Human-Computer Interaction,HCII 2020 - Copenhagen, Denmark
Duration: 19 Jul 202024 Jul 2020

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume12425 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference22nd International Conference on Human-Computer Interaction,HCII 2020
CountryDenmark
CityCopenhagen
Period19/07/2024/07/20

Keywords

  • Behavlets
  • Churn prediction
  • Entropy
  • Player modelling
  • Processing mining

Fingerprint

Dive into the research topics of 'Behavlet Analytics for Player Profiling and Churn Prediction'. Together they form a unique fingerprint.

Cite this