基于协同过滤的银行产品推荐系统建模  

Modeling of Recommendation System for Financial Products Based on Collaborative Filtering

作　　者：李欣璐[1] 刘鲁[1] 

机构地区：[1]北京航空航天大学经济管理学院,北京100083

出　　处：《计算机与数字工程》2007年第9期6-8,66,共4页Computer & Digital Engineering

基　　金：国家自然科学基金项目(编号:70371004)资助

摘　　要：通过分析银行产品推荐系统与一般推荐系统的区别,采用协同过滤算法,设计银行产品推荐系统模型。首先采用决策树、聚类等数据挖掘方法进行客户分类,提高系统伸缩能力和推荐效率;同时从客户和产品两方面对交易明细进行数据分析,避免早期数据冷起动问题。该系统模型最终生成的客户/产品/时间匹配矩阵,有效辅助银行的客户关系管理及市场营销活动。Based on the characteristic of financial products and banking, the paper gives a model of recommendation system for financial products, using collaborative filtering algorithm. Decision tree and clustering are used to analyze customers before recommendation to improve the system efficiency. The analysis both of customers and products avoid the problem of none-match data for new product in the algorithm of collaborative filtering. A matching matrix of customer, financial product and time will be created, which can strengthen the effect of customer relationship management and marketing & sales strategy of banks.

关 键 词：协同过滤 推荐系统 银行产品 客户关系管理 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]