Predictive modeling in credit risk: a survival analysis case

Omoga, Allan Anyona

Predictive modeling in credit risk: a survival analysis case - Nairobi Strathmore University 2017 - viii,39p;

Six survival analysis techniques are accessed by applying the techniques to a dataset consisting of 33,238 active credit facilities from a financial institution operating in Kenya. Namely, the Accelerated Failure Time (AFT) Models, Cox proportional hazard (PH) Model and the Mixture Cure Model (MCM) are considered in the comparisons. Evaluation of the techniques is conducted from a Statistical approach evaluation using the Area under the Curve (AUC) and financial evaluation using the annuity theory. The Cox Proportional Hazard (PH) and the Mixture cure model performs significantly well.
URI


Credit risk modeling

HG3751.O46 2017

© Strathmore University Library
Madaraka Estate, Ole Sangale Road | P. O. Box 59857 - 00200 City Square, Nairobi, Kenya
(+254) (0)703 034 000/200/300 | (+254) (0)20-607498