一种新颖的混合相似度计算模型

摘要

Aiming at the problem of similarity calculation in the traditional neighborhood cooperative filtering recommendation algorithm,a new hybrid similarity calculation model is proposed from the point of view of solving the problem of user interest drift based on the sparseness ofdata.The model is composed of two parts,on the one hand,using the fitting function characterizes the behavior scores and score the preferences of the user itself,on the other hand,the random forest method consider the attributes of users,and two aspects to construct a new hybrid similarity calculation model.The experimental results showed that the prediction accuracy of the proposed algorithm was higher than that of the traditional recommendation algorithm.%传统的最近邻协同过滤推荐算法中相似度计算存在一些问题,如不能刻画变化的用户偏好.在数据稀疏的基础上,从解决用户兴趣漂移问题的角度出发,提出一种新的混合相似度计算模型.该模型由两部分组成:一方面利用了函数拟合刻画了用户自身的评分行为和评分偏好;另一方面采用随机森林方法考虑了用户的属性特征,并综合两方面构建了一种新的混合相似度计算模型.实验结果显示,在不同的数据集规模中,该模型算法的预测精度比传统推荐算法高.