【目的/意義】在利用用戶感興趣資源進行用戶興趣建模中,傳統的資源特征選擇方案未能體現用戶真實興趣,針對這一情況,提出一種基于認知的用戶興趣建模方法,改善個性化推薦效果。【方法/過程】在結合用戶群體認知對資源特征進行識別的基礎上,對用戶感興趣資源進行興趣建模。以電影數據為例,進行個性化推薦實驗,驗證模型效果。【結果/結論】實驗結果顯示,基于認知的用戶興趣建模的推薦準確率明顯高于傳統基于項目的用戶興趣建模方法,該策略可以更準確地描述用戶興趣,提升用戶興趣建模效果。

【Purpose/significance】 The traditional of resource features cannot reflect users’ real interest when using resources that users interested in to construct user profiles modeling. To solve this problem, this paper proposes a user profiles modeling based on cognition to improve the effect of personalized recommendation.【Method/process】On the basis of recognizing the resource characteristics by considering the group cognition, we use resources that they are interested in to express users’ interest, and verifies it by experiment based on movie data.【Result/conclusion】The results show that the effectiveness of user profiles modeling based on cognition is apparently better than item-based user profiles. Thus, it can be concluded that this model can describe users’ interest more accurately, then improve the effect of user profiles modeling.