A System for Credit Appraising - An application of the LogitBoost Algorithm
View/ Open
Date
2011Author
Kirori, Zachary Kamau
Type
ThesisLanguage
enMetadata
Show full item recordAbstract
Mitigation of credit risk is a key aspect of portfolio management in any financial
institution. This is chiefly due to difficulties in uncovering uncertainties in information
provided by credit applicants and also due to lack of reliable automated techniques that
would improve the efficiency of manual underwriting procedures. In this document, we
report on an application of the logistic regression meta learning algorithm in
development of a computer system that could greatly enhance the underwriting
process.
The implementation is based on the java platform to create an interface that can
be used to train a model and use it predictions for credit decisions. The results obtained
prove that such a mechanism can be applied to augment credit appraising processes,
especially where large volumes of applications are to be processed within limited
timeframes.
Publisher
School of Computing and Informatics
Description
MSc