Investigating combinations of machine learning and classification techniques in a game environment

Ruth Edmundson, Richard Danby, Kris Brotherton, Emma Livingstone, Lee Allcock

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

An open world turn based monster battle game was developed in Java using the popular LibGDX game framework applying multiple machine learning algorithms for its mechanics consisting of an ID3 decision tree, perceptron, naïve Bayes classifier and A* pathfinding in an attempt to imitate `machine intelligence'. A tiled map was used as the game area containing multiple AI agents with different personalities that change depending on the difficulty level chosen. The aim of the game focuses on the player defeating each `intelligent machine' non-player character's (NPC) upon interaction with each other, when player and enemy NPC sprites meet a battle screen appears to allow the player and enemy to engage in a turn-based battle with their monsters. When a battle is lost the player loses a life, otherwise they can approach and engage other enemy agents to battle on the map, and thus the game is called `Battle Monsters'.
Original languageEnglish
Title of host publication2016 12th International Conference on Natural Computation, Fuzzy Systems and Knowledge Discovery (ICNC-FSKD)
Place of PublicationPiscataway
PublisherIEEE
Pages1306-1311
ISBN (Print)978-1-5090-4094-0
DOIs
Publication statusPublished - Aug 2016
Externally publishedYes

Fingerprint Dive into the research topics of 'Investigating combinations of machine learning and classification techniques in a game environment'. Together they form a unique fingerprint.

Cite this