Optimizing mobile electronic program guide by adaptive learning

Chengyuan Peng, Petri Vuorimaa

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

Abstract

The Electronic Program Guide (EPG) is the most important service of future mobile digital television platform. In this paper, we present an approach which applying adaptive learning algorithms to the EPG agent to automatically learn optimized channel categorization and TV event list prompts. We apply a learning Classifier System (especially XCS) as a reinforcement learning component to learn policies (a set of rules). We embed usability heuristics of mobile digital television in the reward functions, which are used to achieve the agent's goals. We test this approach and demonstrate that the accuracy of channel classification can be as high as 99% in average.

Original languageEnglish
Title of host publicationProceedings of the Second IASTED International Conference on Communications, Internet, and Information Technology
EditorsM.H. Hamza, M.H. Hamza
PublisherActa Press
Pages814-819
Number of pages6
ISBN (Print)0-889-86398-9
Publication statusPublished - 1 Dec 2003
MoE publication typeA4 Article in a conference publication
Event2nd IASTED International Conference on Communications, Internet, and Information Technology - Scottdale, AZ, United States
Duration: 17 Nov 200319 Nov 2003

Conference

Conference2nd IASTED International Conference on Communications, Internet, and Information Technology
CountryUnited States
CityScottdale, AZ
Period17/11/0319/11/03

    Fingerprint

Keywords

  • Channel Categorization
  • Mobile Digital Television
  • Reinforcement Learning
  • Reward Function
  • XCS

Cite this

Peng, C., & Vuorimaa, P. (2003). Optimizing mobile electronic program guide by adaptive learning. In M. H. Hamza, & M. H. Hamza (Eds.), Proceedings of the Second IASTED International Conference on Communications, Internet, and Information Technology (pp. 814-819). [403-012] Acta Press.