Learning from preferences and selected multimodal features of players

The influence of multimodal sources of input data to the construction of accurate computational models of user preferences is investigated in this paper. The case study presented explores player entertainment preferences of physical game variants incorporating two data modalities. The main findings of the paper reveal the benefit of multiple modalities of input data for the prediction of preferences and highlight the impact of feature selection on the construction of such models. Categories and Subject Descriptors I.2.1 [Artificial Intelligence]: Applications and Expert Systems—Games; I.2.6 [Artificial Intelligence]: Learn

• Type: Conference
• Year: 2009
• Where: ICMI
• Authors: Georgios N. Yannakakis
• Keywords: Accurate Computational Models, Biometrics, ICMI 2009, Input Data, Player Entertainment Preferences