基于Minhash的协同过滤技术在推荐系统中的应用  

Collaborative Filtering Technology Based on Minhash and Its Application in Recommend System

作　　者：刘艾侠[1] 刘丹丹 

机构地区：[1]宝鸡职业技术学院,宝鸡721013 [2]中国科学院大学,北京100039 [3]中国科学院国家授时中心,西安710600

出　　处：《微型电脑应用》2017年第10期67-70,共4页Microcomputer Applications

摘　　要：传统协同过滤的推荐机制应用在大规模数据上时,如果在要保证推荐质量会导致占用运行时间和存储空间的增加。研究分析了Minhash在大规模数据上的降维原理,论证了将minhash应用到协同过滤,设计并实现基于Minhash算法的协同过滤模型。实验结果表明Minhash能在保证推荐质量的前提下很大程度上缩短计算时间和存储空间,能有效地扩展到大规模数据集。The traditional collaborative filtering recommendation mechanism, in order to ensure the quality of recommendation for a set of lavge scale data, will lead to increase run time and memory space. This paper analyzes Minhash dimensionality reduction principles for large-scale data, argues to was Minhash for collaborative filtering, and then designs and implements collaborative filtering algorithm based on Minhash model. The results show that Minbash can largely reduce computing time and storage space under the premise of quality assurance, and can effectively extend to large data sets.

关 键 词：协同过滤 大规模数据集 Minhash 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]