基于用户语义相似性的协同过滤推荐算法

Collaborative Filtering Recommendation Algorithm Based on User Semantic Similarity

为了解决协同过滤推荐中的稀疏性问题,提出一种基于用户语义相似性的协同过滤推荐算法。算法考虑到项目之间内在的语义关系,通过构建领域本体来计算项目之间的语义相似度,并综合项目语义相似度和用户评分数据来度量用户语义相似性。实验结果表明,该算法在用户评分数据极端稀疏的情况下,依然可以获得较高的推荐质量。

In order to resolve the sparsity problem of collaborative filtering recommendation, a collaborative filte- ring recommendation algorithm based on user semantic similarity is presented. In consideration of the inner semantic relationship among items, the algorithm calculates the semantic similarity among items by constructing do main ontology, and combines the item semantic similarity and user rating data to measure the user semantic simi larity. The experiment result shows that this algorithm can effectively improve the recommendation quality in the condition of extreme sparsity of user rating data.