关键词: 上下文感知, 规则生成, 接受度, 前摄推荐, 决策树

Abstract:

 In order to reduce the potential interference with users′ current activities through proactive recommendation and improve the users′ experience, a proactive recommendation model of adapting to users′ changing requirements was proposed. This model concerned with the users′ potential demands for proactive information given by the system, and especially the volatility of those demands. This model adopted an acceptance vector to evaluate the user′s demand for proactive information,  and the decision tree algorithm was  applied to conduct rule-based reasoning by means of the user′s historical context, and then the users′ acceptance vectors for context recommendation was given. The recommendation acceptance vectors could be adjusted according to the changing  context and user′s feedback. The experimental results showed that the proposed  model could adaptively respond to the users′ changing requirements, which was helpful to improve the user experience.

Key words: rule generation, decision trees, acceptance degree, proactive recommendation, contextaware