基于内容预测和项目评分的协同过滤推荐

文中提出了一种基于内容预测和项目评分的协同过滤推荐算法，根据基于内容的推荐计算出用户对未评分项目的评分，在此基础上采用一种基于项目的协同过滤推荐算法计算项目的相似性，随后作出预测。实验结果表明，该算法可以有效解决用户评分数据极端稀疏的情况，同时运用基于项目的相似性度量方法改善了推荐的精确性，显著提高推荐系统的推荐质量。

Collaborative Filtering Recommendation Based on Content and Item Rating Prediction

Traditional similarity measure methods work poor in this situation,which makes the quality of recommendation system decrease dramatically. To address this issue a novel collaborative filtering algorithm based on content and item rating prediction is proposed. This method predicts item ratings that usesr have not rated based on content prediction and then uses item-based collaborative filtering to find similar items and make a prediction. The experiment results suggeste that this method can efficiently improve the extreme sparsity of user rating data,improve accuracy of recommendation using item-based collaborative filtering,and provide better recommendation results than nearest neighborhood collaborative filtering algorithms.