基于网络购物评论的协同过滤推荐算法

Collaborative Filtering Recommendation Algorithm Based on Online Shopping Reviews

提出融合网络购物评论的协同过滤推荐算法,通过挖掘评论信息中产品特征属性,发现用户对不同属性的偏好程度,建立用户产品属性层级偏好模型矩阵,利用改进的余弦相似度计算用户之间的相似度,并预测用户对未购买的产品特征属性的评分得出产品的推荐列表。实验结果表明,提出的算法与常用的推荐算法相比,改善推荐的质量和推荐精度。

Proposes a collaboration filtering recommendation algorithm that integrates online shopping reviews,by mining product attribute attributes in comment information,finds the users'preferences for different attributes,builds the user attribute hierarchies,and calculates the similar.ity between users by using improved cosine similarity,and predicts the user's rating of the product's non-purchased product attribute to de.rive a product's recommended list.The experimental results show that the proposed algorithm improves the recommended quality and rec.ommended accuracy compared with the commonly used recommended algorithms.