融合内容和改进协同过滤的个性化推荐算法

A Personalized Recommendation Algorithm Using Content and Improving Collaborative Filtering

为了提高个性化推荐算法的推荐准确性,提出一种融合内容和改进协同过滤的推荐算法。首先,利用基于内容的过滤方法对未评分的项目进行预测,获得预测评分,从而构建了项目—评分矩阵,用于计算评分的Pearson相关系数。然后,在Pearson相关系数计算中融入项目热门系数,对传统协同过滤方法进行改进,并以此对项目给出最终的评分,从而产生推荐。另外,通过调和加权因子和用户加权因子,使基于内容推荐和协同过滤的评分结果能够更合理的融合。实验结果表明,与当前个性化推荐技术相比,所提算法能够有效解决用户评分数据稀疏的问题,具有更高的推荐精度。

In order to improve the recommendation accuracy of the personalized recommendation algorithm, a recommendation algorithm of fusion content and improving collaborative filtering is proposed. Firstly, the content-based filtering method is used to predict the unrated items to obtain the forecast scores, and build the project-scoring matrix which is used to calculate the Pearson correlation coefficient of the score. Then, the project hotspot coefficient is integrated in the Pearson correlation coefficient calculation, so as to improve the traditional collaborative filtering method. Finally, it gives the final score for the project and produces recommendations. In addition, the score based on content recommendation and collaborative filtering can be more rationally integrated by combining the weighting factor and the user weighting factor. Experimental results show that, compared with the current personalized recommendation technology, this method can effectively solve the problem of sparse user scoring data with higher recommendation accuracy.