摘要
协同过滤推荐是最成功的推荐技术之一,但数据稀疏性问题导致推荐准确度和推荐效率不高。针对这个问题,提出了一种改进的加权Slope one协同过滤推荐算法。计算用户之间的评分相似度,找出每个用户的最近邻;根据最近邻用户评分,使用基于用户的协同过滤和改进的加权Slope one算法的加权评分预测目标用户的未评分项目;给出推荐。实验过程中采用Movie Lens数据集作为测试数据。实验结果表明:与原算法相比,算法提高了预测准确度,有效提高了推荐性能。
Collaborative filtering is one of the most successful recommendation technologies,but the data sparsity results in low recommendation accuracy and poor efficiency. So an improved weighted Slope one and collaborative filtering algorithm is proposed. Based on users' ratings, it calculates the similarity between users, so that to find user' s the nearest neighbors. Based on the score of user' s nearest neighbors, user-based collaborative filtering and weighted Slope one algorithm is used to predict the unknown rating of the target users and to present recommendation results. In the experiment, MovieLens data set is used as test data. The experimental results suggest that the improved algorithm improves prediction accuracy and recommendation performance.
出处
《传感器与微系统》
CSCD
2017年第7期138-141,共4页
Transducer and Microsystem Technologies
基金
国家自然科学基金资助项目(61462053
31300938)