Design and evaluation of adaptive assistance for the selection of movies

This paper describes the application and evaluation of an Adaptive TV Program Agent System with adaptive active and passive assistant functionalities. To evaluate the functionalities of the passive recommendation functions a user-test was conducted to compare different variants of adaptive assistance for the selection of movies. Based on a personal user-profile each participant was offered suggestions that were generated by an Adaptive TV Program Agent System according to different rules. Subjects' task was to choose the movies they would actually intend to watch. To measure the acceptance of these different variants of adaptive assistance it was recorded in log files how many suggestions were accepted per rule. Results show that suggestions based on information about genre, country, year and mood of movies, that appear combined in the profiles reached an acceptance rate of 90 %.