【摘要】 个性化推荐技术为人们处理信息过载问题提供了一种有效的解决方式。Slope One算法是预测评分的推荐算法，通过用户对项目的评分差异来预测评分，再根据预测评分进行推荐。但是，它并未考虑到项目相似度的问题。因此，提出项目相似度的加权Slope One算法，先使用Person相关系数计算出项目相似度，将Person相关系数归一化后与Slope One算法加权结合。最后，在Movielens数据集上进行实验，发现改进后的算法在MAE值上有较好的结果，使推荐更加准确。更多还原

【Abstract】 Personalized recommendation technology provides an effective solution to the problem of information overload. The Slope One algorithm is a predictive rating recommendation algorithm that predicts ratings by the difference in user ratings of items,and then makes recommendations based on the predicted ratings. However, it does not take into account the item similarity problem. Therefore, we propose a weighted Slope One algorithm for item similarity by first calculating the item similarity using the Person correlation coefficient, normalizing the Person correlation coefficient and then combining it with the weighted Slope One algorithm. Finally, experiments are conducted on Movielens dataset and it is found that the improved algorithm has better results on MAE values, which makes the recommendation more accurate.更多还原