A framework of data mining application process for credit scoring

As one of the application areas of data mining, the credit scoring problem is related to theclassification decision. The problem of classification decision has three professionalbackgrounds: statistics, machine learning, and artificial neural networks. There are manycategories of classification algorithms that using two kinds of model generation approaches.Among them five are commonly used in the credit scoring model building and representativefor different categories of classification techniques: Bayesian linear discriminant analysis,instance-based learning, Logistic regression, model trees M5, and Multi-layer propagationnetworks. They have different input requirements, different output forms and have been usedfor the credit scoring problem both practically and academically.The performances of these five model techniques have been compared by manyresearchers. Not any of them is always...

MoreLess

Year of publication:	2002
Authors:	Liu, Yang
Institutions:	Institut für Wirtschaftsinformatik <Göttingen> / Professur für Anwendungssysteme und E-Business
Subject:	Data Mining \| data mining \| Credit Scoring \| Framework <Informatik>

Check Google Scholar |

More access options

In German libraries (KVK)

I need help

1 Introduction
2 Data mining
2.1 The concept of data mining
2.2 The general process of data mining
2.3 The application architecture of data mining
3 Classification decisions
3.1 Two ways of computer-based classification decisions
3.2 The inductive classification
3.3 Various groups of classification techniques
4 Various classification algorithms for credit scoring
4.1 Discriminant Analysis: Bayesian Linear Discriminant Analysis
4.2 Logistic Regression
4.3 Instance-based learning
4.4 Model Trees: M5
4.5 Neural Networks: Multi-layer perceptron
4.6 Comparisons of the introduced algorithms
5 A framework of the data mining application process for credit scoring
5.1 Reasons for a process framework
5.2 The presentation of the general framework
6 Summary and conclusion

Persistent link: https://www.econbiz.de/10005869163