A theoretical analysis of a pattern recognition algorithm for bank failure prediction

Due to the character of the original source materials and the nature of batch digitization, quality control issues may be present in this document. Please report any quality issues you encounter to digital@library.tamu.edu, referencing the URI of the item.Includes bibliographical references.This thesis describes a theoretical analysis and a series of empirical tests of a pattern recognition based Early Warning System for bank failure prediction. The theoretical analysis centers on the binarization, feature selection and feature construction procedures. Alternative methods for the selection of binarization cutpoints are suggested and tested. Finally, empirical comparisons of the prediction accuracy of the Early Warning System against those obtained with Multiple Discriminant Analysis and logit models are described and analyzed. This research shows that the pattern recognition Early Warning System is a powerful, flexible and accurate prediction and classification system