关键词: 知识推荐, 主题多样性, 知识管理, 知识情境, 工程语义

Abstract: To meet the engineers' needs for the accuracy and diversity of knowledge recommendation simultaneously,a Topic-Diversity Knowledge Recommendation (TDKR) method  was proposed to solve engineering problems.In TDKR,a knowledge item network containing information including engineering semantics,knowledge contexts and problem-solving-oriented knowledge relevance was built for knowledge topic discovery.Based on the divided knowledge topic communities,the concept of user-topic professionalism was proposed to improve user-based collaborative filtering method,and the user interest model was built to mine the engineers' diverse knowledge needs.Combining knowledge context information and user interest model,a post-filtering diversity recommendation strategy was proposed to ensure the context-based effectiveness and topic diversity of knowledge in the recommendation list.An engineering case study was undertaken to illustrate the practicability and effectiveness of TDKR.As evidenced in the evaluation,TDKR could ensure the effectiveness of knowledge in a specific context while improving the diversity of knowledge recommendation.

Key words: knowledge recommendation, topic diversity, knowledge management, knowledge context, engineering semantic