摘要
证券客户划分可以帮助证券公司掌握更多的客户特征,但是传统的证券客户划分方法都是基于客户自身属性进行分析并划分的,缺少对证券客户之间联系信息的分析,使得划分结果具有一定的局限性和偏差性。针对这种问题,首先分析了证券客户的属性,提取客户之间行为的联系信息,在此基础上利用联系发现的方法对证券客户进行划分,通过联系假设、联系产生、联系确定等三个阶段得到划分结果。对比实验表明,新方法减少了计算量,提高了划分效率,具有较好的客户划分准确率。
Customers demarcation in security can help security companies hold more features of customs.But traditional methods are all only based on their properties,lacking of analysis of link information between custnmers.According to the above problems, This paper extracts the link infnrmation between customers through the analysis of the attributes of security customers.On this hasis,the paper suggests a new method of customers demarcation in security based on link discovery.The method gets the demarcation result through there basic steps which are link hypothesis,link generation,and link identiflcation.The experiment shows that the new method has a better precision and can reduce the amount of computation.
出处
《计算机工程与应用》
CSCD
北大核心
2009年第18期201-204,共4页
Computer Engineering and Applications
基金
国家高技术研究发展计划(863)No.2007AA04Z116
国家自然科学基金No.70871033~~
关键词
联系发现
相关分析
客户划分
link discovery
eorrelation analysis
customer demarcation