基于领域知识的协同过滤推荐算法

Collaborative Filtering Recommendation Algorithm based on Domain Knowledge

传统协同过滤推荐算法中项目相似度的计算建立在用户评分项目交集之上,没有考虑不同项目之间所存在的语义关系,致使推荐准确率低。基于领域知识进行项目相似度计算的协同过滤算法在用户评分的共同项目很少的情况下仍能给出不错的推荐。实验结果表明,该算法可以有效地解决用户评分数据极端稀疏的问题,提高推荐系统的推荐质量。

Traditional collaborative filtering recommendation algorithm calculates items similarity using the intersection of different user rating items, does not consider the semantic relationship between different Items, results in a low accuracy rate. A novel collaborative filtering algorithms based on domain knowledge can give good results when user common rating items are sparse. The experimental results show that this method can efficiently improve the extreme sparsity of user rating data, and provide better recommendation results.