基于个性化特征的协同过滤推荐算法

Collaborative filtering recommendation algorithm based on personalized feature

针对相似度特征点推荐方法对用户个性化需求匹配度不高的问题,提出基于个性化特征的协同过滤推荐算法。以社会网络为结构模型构建用户信息的评分模型和项目属性模型,采用信任度条件概率分析方法构建可靠性推荐模型,进行个性化特征分析和提取,实现个性化特征需求与项目兴趣点的合理匹配,实现协同过滤推荐,最后通过仿真实验进行测试分析。结果表明,采用该方法进行社会网络项目协同过滤推荐的用户评分高,平均绝对误差和均方根误差小,提升了推荐质量。

Since the recommendation method of the similarity feature point has low matching degree for the user＇s persona- lized demand, a collaborative filtering recommendation algorithm based on personalized feature is put forward. The scoring model and project attribute model of the user information were constructed by taking the social network as the structure model. The trust degree conditional probability analysis method is adopted to construct the reliability recommendation mode. And then the personalized features are analyzed and extracted to match the personalized feature demand and project interest point reasonably, so as to implement the collaborative filtering recommendation. The test analysis was conducted with simulation experiments. The test results show that the method has high user scoring to perform the collaborative filtering recommendation of the social net- work project, the mean absolute error and root mean square error are small, and the recommendation quality is improved.