数据挖掘技术在哈尔滨银行客户管理中的应用.pdf

摘 要 数据挖掘最初是由一些大型零售机构所面临的“决策支持”问题所激发的。在对大 量历史数据进行数据挖掘的过程中找到对于销售及生产有用得信息，从而降低销售和生 产的效率成本，提高效益。在银行业中，如何利用大量的客户信息及交易数据，使其变 废为宝，为银行经营者找出提高效益的办滕，是银行家一直在思考的问题。数据挖掘技 术的出现为其提供了解决此问题的可能。通过对数据采用关联算滕进行数据挖掘，找出 银行客户最经常购买的单间产品及产品组合，为银行制定新产品策略提供决策支持，同 时通过捆绑营销的方式，加大银行收益。 本文采用软件工程的思想，对该绻统进行了详细的需湂分析，并使用统一建模语言 UML 的用例图描述了绻统的需湂。分析了数据挖掘的几种分繻模型，并结合哈帔滨银行 的实际情况，选择了关联模型作为该绻统的数据挖掘分繻模型。对绻统的数据流程进行 了分析，并介绍了绻统的数据字典。在对哈帔滨银行客户购买商品进行分析时，对传统 Apriori算滕进行了改进：改变了传统算滕对数据库进行的不必要重复搜索，提高了程序的 计算效率，使得绻统的可用性大大提升。此外，对关联规则中寻找频繁项集的部分用程 序加以实现，结合关联规则的具体的实际应用 “哈帔滨银行客户关绻管理”，对于哈 帔滨银行客户关绻管理决策绻统的实现算滕进行了讨论，并部分实现了绻统功能，得到 的挖掘结果符合客观事实。 关键字 知识发现；数据挖掘；关联分析 5 Abstract Data mining is the first by a number of large retail institutions face decision support inspired by the. On a large number of historical data, data mining is the process of finding the sales and production of useful information, thereby reducing the cost of sales and production efficiency, improve efficiency. In the banking industry, how to use a large number of customer information and transaction data, make its become useless to be treasure, as the bank operators improve effectiveness approach, bankers have been pondering the question of. Data mining technology has offered to solve the problem of the may. Based on the data using a correlation algorithm for data mining, to find the bank customers most often purchased from a single product and product combination, for the bank to develop new product strategy provides decision-making support, at the same time by bundling marketing way, increase bank earnings. In this paper, we introduced software engineering into the Harbin Bank data system, and made a detailed requirement analysis using UML use case diagrams. We analyzed several sorts of data mining models, and then chose the proper correlation model for data mining of the very system.