一种优化的推荐算法  

An improved recommendation algorithm

作　　者：王永利[1] 何大韧[1] 

机构地区：[1]扬州大学物理科学与技术学院,江苏扬州225002

出　　处：《扬州大学学报（自然科学版）》2009年第4期18-21,共4页Journal of Yangzhou University：Natural Science Edition

基　　金：国家自然科学基金重点资助项目(10635040);面上资助项目(70671089)

摘　　要：对个性化推荐系统算法进行改进,首先,不仅考虑用户所选的商品,而且考虑用户的打分,从而将资源分配法扩展为含权资源分配;其次,考虑用户的固有相似性.把这两方面相结合,发展了新的算法.数值试验表明,改进后的方法显著提高了推荐的精度和个性化程度.This paper presents an improved algorithm for personal recommendation. It takes not only the merchandise selected by users into account, but also the ratings given by the users. Then, it considers the inherent similarities between the users. Combining these two factors together, a new algorithm is developed. The numerical simulation results show that the presented algorithm can obviously enhance the accuracy and personality of the recommendation.

关 键 词：个性化推荐 资源分配 固有相似性 

分 类 号：N941.4[自然科学总论—系统科学] O414.2[理学—理论物理]