Currently collaborative filtering is the most successful and widely used recommendation technology in recommender systems. However, with the development of E-commerce, the magnitudes of users and commodities grow rapidly, which results in the extreme sparsity of user rating data. The method of searching for nearest neighbors in traditional collaborative filtering algorithm works poor in this situation, which makes the quality of the recommender systems decrease dramatically. To address this issue, a collaborative filtering recommendation algorithm based on item-class preference is proposed. The proposed algorithm first finds out a set of candidate neighbors who are similar to the active user in item-class preference. The candidate neighbors have similar interest and more co-rated items with the active user. Then the algorithm identifies some nearest neighbors in the candidate neighbor set, which eliminates the interference of the users who have few co-rated items with the active user, and enhances the accuracy of searching for nearest neighbors. The experimental results show that the proposed algorithm can efficiently improve recommendation quality.%协同过滤是推荐系统中广泛使用的最成功的推荐技术,但是随着系统中用户数目和商品数目的不断增加,整个商品空间上的用户评分数据极端稀疏,传统协同过滤算法的最近邻搜寻方式存在很大不足,导致推荐质量急剧下降.针对这一问题,本文提出了一种基于项类偏好的协同过滤推荐算法.首先为目标用户找出一组项类偏好一致的候选邻居,候选邻居与目标用户兴趣相近,共同评分较多,在候选邻居中搜寻最近邻,可以排除共同评分较少用户的干扰,从整体上提高最近邻搜寻的准确性.实验结果表明,该算法能有效提高推荐质量.
展开▼
