论文部分内容阅读
如何准确高效地提供给用户需要的信息,是信息推荐研究的核心。提出一种推荐机制——基于信用矩阵的推荐机制,借鉴协同推荐的实现思想,引入一个信用矩阵,在向目标用户推荐时,不是基于最近邻产生推荐,而是基于用户之间的信用产生对目标用户的推荐。在MovieLens上的实验证明,这种算法在预测精度上较传统的推荐算法有很大的提高。
How to accurately and efficiently provide users with the information they need is the core of information recommendation research. This paper proposes a recommendation mechanism based on the credit matrix recommendation mechanism, draws on the idea of collaborative recommendation and introduces a credit matrix. When recommending to target users, not based on the nearest neighbor to generate recommendations, but based on the credit generation between users Target user’s recommendation. Experiments on MovieLens prove that the proposed algorithm has much higher prediction accuracy than the traditional recommended algorithm.