协同过滤算法是网上推荐系统最常用的算法,但是传统的协同过滤算法很难解决数据稀疏、冷启动、用户兴趣变化等问题.本文提出基于信任与用户兴趣变化的协同过滤改进方法.该方法将信任引入到传统协同过滤算法中,构建用户信任模型,用信任的传递特性为用户匹配更多邻居用户,从而可以在一定程度上缓解数据稀疏性等问题.随着时间的变化,用户的兴趣也会发生变化,本文利用时间遗忘函数来模拟用户的兴趣变化.本算法综合用户相似度、用户信任度及用户兴趣变化,为目标用户推荐项目.最后利用数据实验验证本方法的有效性.%The collaborative filtering algorithm is the mostly applied algorithm in the internet recommendation systems,but it's not effective in solving the problems of data sparsity,cold start and the change of user's interest.This paper proposes a collaborative filtering algorithm based on trust and the change of user's interest.The trust is introduced into the traditional collaborative filtering algorithm,build user trust model,use trust transfer characteristic for the user to match more neighbors,which can to some extent alleviate the problems of data sparsity and cold start users.The user's interest will change with times,original score data will not be able to express the user's current interest.In this paper,the forgetting function of time is proposed to simulate the change of user's interest.Taking the user similarity,user trust and the change of user's interests into consideration,we construct the new method to recommend items to the target user.Finally,experimental schemes are used to verify the effectiveness of our method.
展开▼
