欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
为解决传统协同过滤推荐算法相似度矩阵不能局部更新的问题,提出了一种基于增量更新的协同过滤推荐算法。该算法首先根据用户评分数据构建用户相异度矩阵,然后选取与目标用户相异度较小且同现次数较多的若干用户作为目标用户最近邻居并产生推荐。算法可以对相异度矩阵进行在线局部更新,无须离线导入全部数据重新计算,从而实现了算法的增量更新,使算法具备了良好的扩展性。进一步实验表明,基于增量更新的协同过滤算法具有很高的推荐准确性。
英文摘要：
In order to solve the problem that the similarity matrix of the traditional collaborative filtering recommendation algorithm cannot update partial. A collaborative filtering recommendation algorithm based on incremental updating is proposed. First,the algorithm constructs user- deviation matrix according to the user rating data. And then it selects a number of users whose user- deviation with target user is small and the co-occurrence times is high as nearest neighbor and generates recommendations. This algorithm can update partial for the deviation matrix online. It doesn＇t have to import all the data to recalculate offline. It implements the incremental updating of the algorithm,which makes it has a good scalability. The experiment result shows that the collaborative filtering algorithm proposed by this paper has very high recommendation accuracy.
同期刊论文项目
同项目期刊论文
期刊信息
位置：