一种融合协同过滤和用户属性过滤的混合推荐算法

A hybrid recommendation algorithm based on collaborative filtering and user attribute filtering

传统的推荐算法向用户进行推荐时一般以用户评分矩阵作为基础,向用户推荐相应的内容,但评分矩阵数据不充分时,该推荐算法准确性难以得到保障。本文中所述的融合协同过滤和用户属性过滤的混合推荐算法,提出时间热度的计算方法并对Pearson相关系数进行改进,建立用户属性相似度模型,对邻居用户进行过滤,由最终票选得到的可信邻居用户向当前匹配用户推荐。经过的系列实验的结果表明,本文中提出的融合协同过滤和用户属性过滤的混合推荐算法较之前经典的系统过滤算法有更好的效果。

The traditional Collaborative Filtering（CF）recommendation algorithm is based on the user scoring matrix to recommend to the user. There is a problem that the recommendation information is inaccurate due to sparse data. Accordingly we propose a hybrid recommendation algorithm which combines cooperative filtering and user attribute filtering. In this paper,we first propose the method of calculating the time heat and improve the Pearson correlation coefficient algorithm. And then establish the user attribute similarity model. Filtering the neighbor user,and recommending by trusted neighbors user finally obtained to the current user. The experimental results show that the hybrid recommendation algorithm proposed in this paper has better effect than the traditional system filtering algorithm.