Credit Intelligence & Modelling

Abstract

This book, “Forest Paths” for short, started as a detailed guide for the construction of predictive models for credit and other risk assessment, for use in big-bank retail lending. It became a textbook covering credit processes (from marketing through to fraud), bureau and rating agencies, and various tools. Included are detailed histories (economics, statistics, social science}, which much referencing. It is unique in the field, with chatpers’-end questions. The primary target market is corporate and academic, but much would be of interest to a broader audience. There are eight modules: 1) an introduction to credit risk assessment and predictive modelling; 2) micro-histories of credit, credit intelligence, credit scoring, plus industrial revolutions, economic ups and downs, and both personal registration and identification; 4) mathematical and statistical tools used to develop and assess predictive models; 5) project management and data assembly; 6) data preparation from sampling to reject inference; 7) model training through to implementation; and 8) appendices, including an extensive glossary, bibliography, and index. Although the focus is credit risk, especially in the retail consumer and small-business segments, many concepts are common across disciplines as diverse as psychology, biology, engineering, and computer science, whether academic research or practical use. It also covers issues relating to the use of machine learning for credit risk assessment. Most of the focus is on traditional modelling techniques, but the increasing use of machine learning is recognised, as are its limitations. It is hoped that the contents will inform both camps.