Abstract: With the rapid development of the Internet, a large variety of information has exploded, and resulting in information overload. Now, by analyzing a large number of available information, recommender systems can help users find things they are interested. And collaborative filtering is the most widely used approach in the recommendation systems. However, its accuracy of recommendation is still needed to be improved. In this paper, a novel effective collaborative filtering algorithm based on segmented and improved PCC is proposed to improve the accuracy of the recommendation system. The number of co-items and the PCC threshold is used to calculate and improve the results of the PCC in this method. Finally, the experimental results show that the proposed method is better than other traditional methods.

Key words: recommender system, collaborative filtering, similarity, segmentation

摘要： 随着互联网的快速发展，大量各式各样的信息呈爆发式增长，导致了信息过载。如今，推荐系统可以通过分析大量的可用信息帮助用户找到他们感兴趣的对象。其中，协同过滤算法是推荐系统中使用得最广泛的推荐算法。但是，协同过滤推荐算法在推荐的准确度上还有待改进。提出了一种基于多分段改进PCC的协同过滤推荐算法，用于提高推荐系统的准确度。提出的方法将根据用户公共项目数和PCC阈值，对PCC算法进行分段计算并改进结果。最后的实验结果表明，该方法的推荐效果要优于其他传统的推荐方法。

关键词: 推荐系统, 协同过滤, 相似度, 分段