Guangli Nie

Credit card churn forecasting by logistic regression and decision tree

In this paper, two data mining algorithms are applied to build a churn prediction model using credit card data collected from a real Chinese bank. The contribution of four variable categories: customer information, card information, risk information, and transaction activity information are examined. The paper analyzes a process of dealing with variables when data is obtained from a database instead of a survey. Instead of considering the all 135 variables into the model directly, it selects the certain variables from the perspective of not only correlation but also economic sense. In addition to the accuracy of analytic results, the paper designs a misclassification cost measurement by taking the two types error and the economic sense into account, which is more suitable to evaluate the credit card churn prediction model. The algorithms used in this study include logistic regression and decision tree which are proven mature and powerful classification algorithms. The test result shows that regression performs a little better than decision tree.

Decision analysis of data mining project based on Bayesian risk

Data mining, an efficient method of business intelligence, is a process to extract knowledge from large scale data. As the augment of the size of enterprise and the data, data mining as a way to make use of the data become more and more necessary. But now most of the literatures only focus on the algorithm itself. Few literatures research what qualification to fulfill before the decision doing data mining from the perspective of the company manager. This paper discusses the factors affect the data mining project. Based on the Bayesian risk, we build a model taking the risk attitude of the top executive in account to help them make decision whether to do data mining or not.

Finding the Hidden Pattern of Credit Card Holder's Churn: A Case of China

In this paper, we propose a framework of the whole process of churn prediction of credit card holder. In order to make the knowledge extracted from data mining more executable, we take the execution of the model into account during the whole process from variable designing to model understanding. Using the Logistic regression, we build a model based on the data of more than 5000 credit card holders. The tests of model perform very well.

A Survey of Interestingness Measures for Association Rules

Association mining can generate large quantity of rules, most of which are not interesting to the user. Interestingness measures are used to find the truly interesting rules. This paper presents a review of the available literature on the various interestingness measures, which generally can be divided into two categories: objective measures based on the statistical strengths or properties of the discovered rules, and subjective measures which are derived from the user’s beliefs or expectations of their particular problem domain. We sum up twelve measure criteria which are concerned by many researchers and evaluate the strengths and weaknesses of the two categories of measures. At last, we pointed out that the combination of objective and subjective measures would be a possible research direction.

Credit card customer analysis based on panel data clustering

Abstract In this paper, we propose a new distance measurement which can be used in panel data clustering. The distance as we designed can be calculated with weight and without weight. If users put more attention on recent data, a heavier weight can be assigned to the recent data. We use real panel data of a commercial bank’s credit card to examine the performance of our new distance measurement. The results show that our distance measurement can reflect the information of different periods and panel data can be used to cluster to find new knowledge. This study discovers different knowledge structure from the traditional econometrics analysis with the help of data mining algorithms.

The Analysis on the Customers Churn of Charge Email Based on Data Mining Take One Internet Company for Example

The email has profoundly affected our ways of life. In the year 2001, many Web sites begin operating the charge emails in china. Recently the foreign Internet company serves free email with large storage capability, which threats the existence of charge email. The churn of email users is serious. This paper studied the churn of the customer using the way of data mining based on the background introduced. There are a lot models and study papers on data mining. This is the same on the churn. But there is few study paper about the churn of charge email based on data mining. So this is the innovation of this paper. Then we analyzed the data after processing the data by way of data cleaning and data integration based on the method of data mining. The data includes the log-data which contain the information of using the email and the database-data which was provided when the client applied the email. The model which the author used to train the data is decision tree. The models the author get after the operation implicate the feature of the churn before they stop using the email. By reasonable explanation of the model, we can help the enterprise to improve the customer relationship management (CRM) and redeem the customers who would leave the email

A New Kernel-Based Classification Algorithm

A new kernel-based learning algorithm called kernel affine subspace nearest point (KASNP) approach is proposed in this paper. Inspired by the geometrical explanation of Support Vector Machines (SVMs) and its nearest point problem in convex hulls, we extend the convex hull of each class to its corresponding affine subspace in high dimensional space induced by kernel. In two class affine subspaces, KASNP finds the nearest points and then constructs a separating hyperplane, which bisects the line segment joining them. The nearest point problem of KASNP is only an unconstrained optimal problem whose solution can be directly computed. Compared with SVM, KASNP avoids solving convex quadratic programming. Experiments on two-spiral dataset, two UCI credit datasets, and face recognition datasets show that our proposed KASNP is effective for data classification.

Retail Exposures Credit Scoring Models for Chinese Commercial Banks

This paper firstly discussed several credit scoring models and their development history, then designed the target system of individual credit scoring with individual housing loans data of a stated-owned commercial bank and logistic method, and established an individual credit scoring model including testing. Finally, the paper discussed the application of the individual credit scoring model in consumer credit domain, and brought forward corresponding conclusions and policies.

Find Intelligent Knowledge by Second-Order Mining: Three Cases from China

This paper discusses the second-order mining of the results of data mining. There is still gap between the knowledge which can direct the operation of company and the knowledge got from data mining. We take the knowledge from data mining as primary knowledge and the knowledge from second-order data mining as intelligent knowledge. We discuss the importance of intelligent knowledge and the way to find intelligent knowledge from primary knowledge via second-order mining process in this study. Three cases from China are used to demonstrate the methods to finish second mining. These cases are related to central bank credit scoring, credit card churn prediction, and email service fields respectively.

Decision support for target country selection of future generation sovereign wealth funds: Hedging the country industry risk

Abstract This paper addresses the challenging problem of selecting target country for future Sovereign Wealth Funds’ (SWFs) asset allocation to hedge the industry risk, which is rarely studied in the field. The target country selection includes which country and how much to invest to obtain the return objective and minimize the risk of these funds. In terms of the industrial perspective, the home country as the investor should consider SWF as part of its budget to make decision in long term. In order to control the risk, this paper measures the similarity between the home and the recipient country of SWF investment. The industrial risk of SWFs’ recipient country is also taken into consideration which is measured by concentration ratio. Based on an analytical process of target country selection, the paper finds that Kazakhstan, India, Australia, Greece, Spain, United States, Austria, Portugal, Peru, Netherlands are the top 10 countries that China should consider as its investment priorities.