Hybrid multivariate adaptive regression splines credit risk assessment model considering business cycle Cover Image

Hybrid multivariate adaptive regression splines credit risk assessment model considering business cycle
Hybrid multivariate adaptive regression splines credit risk assessment model considering business cycle

Author(s): Ričardas Mileris
Subject(s): Business Economy / Management, Economic policy, Methodology and research technology
Published by: Kauno Technologijos Universitetas
Keywords: Bankruptcy; business cycle; credit risk; statistical models;

Summary/Abstract: The paper presents hybrid statistical enterprises classification model that combines multivariate adaptive regression splines and logistic regression methods. According to 11 financial ratios the model allows to predict the default of a company in 1 year. The hybrid approach allowed to increase the sensitivity of model from 90% to 92% while the overall classification accuracy remained the same 96,92% as in logistic regression model. In addition this model considers the business cycle effect on different industry sectors credit risk. Analyzing 5 factors of industry sectors performance the model highlighted the sectors that are very sensitive to the changes in macroeconomic conditions. That can help banks to make lending decisions aiming to reduce the possible loss.

  • Issue Year: 2013
  • Issue No: 05
  • Page Range: 86-94
  • Page Count: 9
  • Language: English
Toggle Accessibility Mode