一种基于支持向量机回归的推荐算法  

A recommendation algorithm based on Support Vector Regression

作　　者：王宏宇[1] 糜仲春[1] 梁晓艳[1] 叶跃祥[1] 

机构地区：[1]中国科学技术大学管理学院,合肥230026

出　　处：《中国科学院研究生院学报》2007年第6期742-748,共7页Journal of the Graduate School of the Chinese Academy of Sciences

摘　　要：随着电子商务的迅速发展,推荐系统与算法已经成为理论研究的热点.支持向量机是一种强大的分类工具,由其衍生出的支持向量机回归方法能很好地解决非线性回归问题.以电影推荐为例,引入支持向量机回归方法来分析项目的内容,构建用户模型,进而给出推荐.实验结果和理论分析表明,这种推荐算法与传统协同过滤算法相比,能够明显提高推荐精度,并显著缩短了推荐所需时间;在大样本量情况下也能同样高效.Recommender systems and recommendation algorithm has become one of the hotspots of data mining research, with the rapid boosting of e-commerce. Support Vector Regression （SVR） algorithm has been introduced to construct a content-based recommend approach. First, the contents of rated items are analyzed with SVR to build regression model of user profiles for active users. Then use the user profiles to give recommendations. Experimental results on the EachMovie dataset show that the proposed approach has better recommend performance and less time spending than the conventional collaborative filtering approach.

关 键 词：推荐系统 支持向量机回归 基于内容的推荐 

分 类 号：P181[天文地球—天文学]