Exploring learning techniques based on decision trees and their performance in platform games
Metadata
Show full item recordcomunitat-uji-handle:10234/158176
comunitat-uji-handle2:10234/71324
comunitat-uji-handle3:10234/169451
comunitat-uji-handle4:
TFG-TFMMetadata
Title
Exploring learning techniques based on decision trees and their performance in platform gamesAuthor (s)
Tutor/Supervisor; University.Department
Sanz Valero, Pedro José; Universitat Jaume I. Departament d'Enginyeria i Ciència dels ComputadorsDate
2020-07-09Publisher
Universitat Jaume IAbstract
This document presents the Final Degree Work of the Bachelor’s Degree in Video Game
Design and Development. The work consists of the study and implementation of machine
learning techniques based on decision trees. ... [+]
This document presents the Final Degree Work of the Bachelor’s Degree in Video Game
Design and Development. The work consists of the study and implementation of machine
learning techniques based on decision trees. The focus is set on Quinlan’s Inductive
Decision Tree algorithm (ID3) and its extension, the Incremental Decision Tree learning
algorithm (ID4).
The learning methods are applied to the classic Super Mario Bros. The artificial
intelligence agents are implemented and trained within the Mario AI Framework . This is a
framework for using AI methods with a version of Super Mario Bros. The framework includes
features such as level generators, observation grid, and already implemented playing
agents.
In order to demonstrate the reliability and feasibility of the system, some tests have been
carried out as an experimental validation. These preliminary results showcase the pros and
cons of the applied learning approach and open the door to continue exploring learning
techniques in other videogame contexts. [-]
Subject
Description
Treball final de Grau en Disseny i Desenvolupament de Videojocs. Codi: VJ1241. Curs acadèmic: 2019/2020
Type
info:eu-repo/semantics/bachelorThesisRights
info:eu-repo/semantics/openAccess
This item appears in the folowing collection(s)
The following license files are associated with this item: