ID3分类算法在银行客户流失中的应用研究

决策树已被成功应用到许多分类问题上,其中ID3是决策树学习的典型算法.文中就该算法在银行客户流失中的应用做了实例研究.叙述了ID3分类算法的原理及其实现算法,并分析了银行客户流失的原因和分类,以一个具体案例详细讲解了ID3分类算法在银行客户流失分析的具体应用流程,包括:数据采样、数据分析、建立模型和模型解释.文中实现ID3算法并作用于银行数据得到一个银行客户流失模型,通过提取模型中的规则对银行预测客户流失特征具有一定的辅助作用.     

一种基于电信客户流失预警的解决方案

本文通过对电信运营商海量数据的统一整合,并对客户的基本属性、呼叫行为、缴费情况、客户服务投诉情况等数据深入研究,分析出已流失或有流失趋势客户的行为特征,建立了客户流失模型,预测具有流失倾向的客户,进行预警.并分析流失原因,针对可能流失的客户设计挽留方案,对挽留方案进行实施、跟踪和评估,形成流失分析和管理的闭环流.     

人工数据合成法下的通讯客户预警模型

如何帮助企业提前识别高风险流失客户,已成为许多管理者关心的问题.许多数据挖掘方法用于通讯客户流失案例中时,存在因变量的分布不均匀导致算法精度下降的问题.文章采用人工数据合成法来解决该问题,提出四种客户流失预警模型:GLM-logistic回归模型,GAM-logistic回归模型,Sem-parameter GAM-logistic回归模型和随机森林模型.以AUC和覆盖率-捕获率作为评价指标进行比较,构建出最合适该案例的Sem-parameter GAM-logistic预警模型,以帮助企业减少不必要的客户流失及由此带来的企业损失.     

C4.5算法在移动通信行业客户流失分析中的应用

稳定客户和吸引客户是移动通信企业提高竞争力的关键.基于大量实验数据将数据挖掘的决策树方法引入移动通信行业客户流失分析中,通过对数据的预处理,利用C4.5算法创建决策树,通过测试流失的与未流失的客户,平均正确识别率为91.6%.决策树体现的规则与经验基本一致,为移动通信企业建立客户流失的预警机制提供了决策支持.     

基于人工神经网络的银行客户流失分析模型

针对目前中国建设银行存在的客户流失问题,利用BP人工神经网络网络稳定、学习能力强的特点,通过输入变量和输出变量之间的相关性分析,建立银行客户流失分析模型,以此获取即将流失的客户,以便银行做出经营决策,挽留有关用户,确保银行效益不受影响。实验证明,此模型能够很好的对银行客户流失进行预测分析。     

基于移动社交用户的流失预警研究

引言 截止2015年2季度手机用户数量突破6.5亿.研究表明,流失一个已有的用户代价是新发展一个用户代价的5倍;用户保持率提高5%,利润将会提升 25%.近年来用户流失预警问题已经在国内外企业和学术界受到越来越多的关注,相关的研究方法广泛应用在电信、银行等相关行业.     

数据挖掘在信用卡客户流失分析中的应用

银行信用卡业务属于高收益、高风险的业务,如何实现对信用卡的客户流失控制是发卡银行迫切需要解决的问题。目前,随着银行积累了大量的数据,并建立了数据仓库,使得采用数据挖掘技术来实现信用卡客户流失分析成为了可能。本文提出了银行信用卡领域内基于数据挖掘的决策分析流程:包括数据准备,数据理解和商业理解阶段,帮助信用卡业务部门分析和控制客户流失。     

基于数据仓库的客户挽留系统

随着移动通信运营市场的竞争日趋激烈,移动电话客户"大进大出",导致客户离网率居高不下,造成营销资源的大量浪费。企业为了保持市场份额和运营效益,通过系统支撑手段加强客户流失挽留工作。锁定中高端客户,通过聚焦客户关怀、业务维系、流失预警挽留等重要手段,有针对性地开展服务营销工作,有效延长客户生命周期,保留存量市场,从而节约营销成本。     

基于数据挖掘的银行客户细分模型

随着银行客户的增多,如何对客户进行分类,制定有针对性的营销策略,保留住优质客户,是银行客户关系管理的重要内容。本文利用X-means算法建立银行客户细分模型,为银行决策者提供科学的决策支持。     

证券公司客户综合分析系统的设计与实现

介绍基于数据仓库与数据挖掘技术的证券公司客户综合分析系统的设计与实现,其中着重介绍了系统的设计原则、设计思想以及有证券特色的数据挖掘模型及其应用等重要内容。用k-Means聚类方法构建了客户偏好细分模型,将客户有效划分为8群;利用决策树及Logistic回归相结合构建了客户流失预警模型,结果表明该模型对客户流失捕获率有很大提升。     

Resuscitating privacy-preserving mobile payment with customer in complete control

Credit/debit card payment transactions do not protect the privacy of the customer. Once the card is handed over to the merchant for payment processing, customers are “no longer in control” on how their card details and money are handled. This leads to card fraud, identity theft, and customer profiling. Therefore, for those customers who value their privacy and security of their payment transactions, this paper proposes a choice—an alternate mobile payment model called “Pre-Paid Mobile HTTPS-based Payment model”. In our proposed payment model, the customer obtains the merchant’s bank account information and then instructs his/her bank to transfer the money to the merchant’s bank account. We utilize near field communication (NFC) protocol to obtain the merchant’s bank account information into the customer’s NFC-enabled smartphone. We also use partially blind signature scheme to hide the customers’ identity from the bank. As a result, our payment model provides the customer with complete control on his/her payments and privacy protection from both the bank and the merchant. We emulated our proposed mobile payment model using Android SDK 2.1 platform and analyzed its execution time.     

基于决策树方法的银行客户关系管理的研究和应用

数据挖掘中的决策树算法在银行业中有很重要的价值。决策树技术应用于银行业中,可以通过对特定的客户背景信息的分析,预测该客户所属的客户类别,从而采取相应的经营策略,这样既可以提高银行服务的服务水平,开发客户资源,避免客户流失,又能够节约资源,利用最小的投入,获得较大的收益。在银行贷款业务中,判断贷款对象是否有风险,贷款方案是否可行,将客户按照银行的实际需求进行分类,这些问题通过决策树算法都可以解决。     

数据挖掘在银行零售业的应用

采用数据挖掘手段,基于某银行零售业的数据,分析了客户的投资偏好。采用CART决策树进行特征筛选,发现客户群体年龄大于30岁,资产处于5万以上且工作稳定的保守型客户更倾向于购买银行基金产品。此外,还构建了逻辑回归模型对客户购买基金的概率进行预测。结果表明,通过数据挖掘相关方法所筛选得到的客户群体有更高的购买概率,因此极大地提高了银行从业人员的工作效率。     

基于RFM模型的随机森林算法对民航客户的流失分析

近几年,随着航空市场的快速发展,对于航空公司而言,如何在增加市场占有率的同时,对客户的流失进行有效的控制也刻不容缓.基于随机森林算法,根据航空客户数据,建立流失预测模型,对客户是否已流失进行预测研究,将传统的RFM客户价值模型进行改进,结合随机森林算法对客户流失进行预测.实验结果表明,基于RFM模型的随机森林算法构建的...     

An integrated data mining and behavioral scoring model for analyzing bank customers

Analyzing bank databases for customer behavior management is difficult since bank databases are multi-dimensional, comprised of monthly account records and daily transaction records. This study proposes an integrated data mining and behavioral scoring model to manage existing credit card customers in a bank. A self-organizing map neural network was used to identify groups of customers based on repayment behavior and recency, frequency, monetary behavioral scoring predicators. It also classified bank customers into three major profitable groups of customers. The resulting groups of customers were then profiled by customer's feature attributes determined using an Apriori association rule inducer. This study demonstrates that identifying customers by a behavioral scoring model is helpful characteristics of customer and facilitates marketing strategy development.     

基于并行C4.5的铁路零散白货客户流失预测研究

为了提高铁路零散白货客户流失预测的准确性和高效性,根据铁路零散白货客户的流失特征,提出了基于CDL模型的客户流失识别方法,在此基础上,针对数据量大的问题,提出了基于Hadoop并行框架的C4.5决策树客户流失预测模型。通过仿真实验,证明该模型具有较好的准确性和预测能力,并且随着样本数量的增加,Hadoop并行框架的效率得到了明显的提升,且不影响客户流失预测模型的准确性和预测能力。     

Pricing and promotion strategies of an online shop based on customer segmentation and multiple objective decision making

The advent of the Internet and web technologies has enabled the prosperity of virtual stores, which greatly reduce customers’ search costs and retailers’ overhead. However, the furious competition between online shops makes it difficult for them to generate profits. This study attempts to establish pricing and promotion strategies for online shops to enhance their profitability. The pricing decision is based on the concept of customer relationship management, where a greater margin of price concession is given to customers who are more valuable to the shop. The process of our approach is: clustering customers into different classes based on their RFM data, computing and presenting the list prices of products to customers according to their classes, allowing customers to bargain over the price and offering conceded prices which are computed based on customer classes and a multi-objective decision making model, and finally providing promotion options to customers to reinforce their purchase inclination. The proposed approach is implemented at an online shop of a computer peripherals retailer. Transaction data before and after the implementation are collected and compared to assess the performance of the proposed approach.     

An incentive-based electronic payment scheme for digital content transactions over the Internet

Digital content transactions through e-commerce will grow tremendous in the coming years. Well-designed electronic payment schemes and high-quality digital contents are two critical successful factors. This paper proposes an incentive-based electronic payment scheme, which can ensure both important properties of fair exchange and customer anonymity in e-commerce transactions and enhance authors’ motivation to create digital contents. The proposed payment scheme is based on cryptographic techniques. Besides, it adopts a mechanism called “the apportionment contract of sales revenue”, which records payees’ apportionment amount. The bank can immediately apportion the sales revenue according to this contract when customers complete payments. By scrupulous analyses, the proposed payment scheme can achieve fair exchange, customer anonymity, and payment security. This paper also compares the proposed payment scheme to related schemes. The comparison result shows that the proposed payment scheme has the best characteristics on the following viewpoints: incentive-based payment, fair exchange, customer anonymity, the denomination of payment token, and the number of payment token for a transaction.     

Extracting knowledge of customers’ preferences in massively multiplayer online role playing games

Due to fierce competition in game markets, to identify customers’ true needs is one of the crucial factors in online game industry. Traditionally, game producers heavily rely on game testers, who are primarily responsible for analyzing computer games, finding software defects and being a part of quality control process, to achieve this goal. But, it is not often reliable. To ensure the investment can be returned, game producers need an effective approach to discover frequently shifted customer preferences in time. Recently, Kano model and data mining techniques have been successfully applied to recognize customers’ preferences and implement customer relationship management tasks, respectively. However, in traditional Kano analysis, only basically statistical analysis techniques are used, and they are insufficient to provide advanced knowledge to enterprisers. Therefore, in order to discover the relationship between/among quality elements in Kano model and to extract knowledge related to customer preferences, this study proposes a knowledge acquisition scheme that integrates several data mining techniques including association rule discovery, decision tree, and self-organizing map neural network, into traditional Kano model. An actual case of customer satisfaction survey regarding massively multiplayer online role playing game has been provided to demonstrate the effectiveness of our proposed scheme.     

Analyzing customer behavior from shopping path data using operation edit distance

Radio frequency identification (RFID) technology has been successfully applied to gather customers’ shopping habits from their motion paths and other behavioral data. The customers’ behavioral data can be used for marketing purposes, such as improving the store layout or optimizing targeted promotions to specific customers. Some data mining techniques, such as clustering algorithms can be used to discover customers’ hidden behaviors from their shopping paths. However, shopping path data has peculiar challenges, including variable length, sequential data, and the need for a special distance measure. Due to these challenges, traditional clustering algorithms cannot be applied to shopping path data. In this paper, we analyze customer behavior from their shopping path data by using a clustering algorithm. We propose a new distance measure for shopping path data, called the Operation edit distance, to solve the aforementioned problems. The proposed distance method enables the RFID customer shopping path data to be processed effectively using clustering algorithms. We have collected a real-world shopping path data from a retail store and applied our method to the dataset. The proposed method effectively determined customers’ shopping patterns from the data.