客户海量兴趣数据分类的推荐系统优化仿真  

Optimization Simulation of Recommendation System in Data Classification of Customer Massive Interest

作　　者：李兰[1] 刘洋[1] 邵明文[1] 

机构地区：[1]青岛理工大学计算机工程学院,山东青岛266071

出　　处：《计算机仿真》2014年第9期449-453,共5页Computer Simulation

基　　金：国家自然科学基金(61173181)

摘　　要：对客户兴趣商品信息数据进行有效推荐,能够最大程度地服务消费者。商品种类和数量快速增长,大量有用商品信息会被淹没在海量商品信息中。传统的客户兴趣推荐系统在进行客户兴趣确认时,受到海量信息的干扰,使得兴趣商品信息数据确认过程耗时巨大,效率较低。提出采用关联决策树算法的客户海量兴趣数据分类推荐系统。在推荐系统中,针对海量数据进行关联决策计算,得到所有数据之间的关联性,根据上述数据关联性,建立关联决策树,获取客户兴趣数据推荐目标。最后进行仿真,结果表明,采用改进方法能够减少商品信息确认时间,取得了良好的效果。Traditional recommender systems for customer interest are affected by huge amounts of information, which makes that confirmation process of commodity information data take long time and low efficiency. A decision tree algorithm is proposed based on the correlation of the customer massive interest in data classification system. In this recommendation system, correlative decision for huge amounts of data is calculated, and the correlation of all the data is obtained. Based on the data correlation, correlation decision tree is established, and the data recommended target of customer interest is obtain. Simulation results show that the improved method can reduce the confirm time of commodity information, and obtain good effects.

关 键 词：数据分类 推荐系统 关联决策树 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]