Deriving Rules for Forecasting Air Carrier Financial Stress and Insolvency: A Genetic Algorithm Approach

Authors

  • Sergio Davalos
  • Richard Gritta
  • Bahram Adrangi

DOI:

https://doi.org/10.5399/osu/jtrf.46.2.1031

Abstract

Statistical and artificial intelligence methods have successfully classified organizational solvency, but are limited in terms of generalization, knowledge on how a conclusion was reached, convergence to a local optima, or inconsistent results. Issues such as dimensionality reduction and feature selection can also affect a model's performance. This research explores the use of the genetic algorithm that has the advantages of the artificial neural network but without its limitations. The genetic algorithm model resulted in a set of easy to understand, if-then rules that were used to assess U.S. air carrier solvency with a 94% accuracy.

Downloads

Published

2010-10-14

Issue

Section

Articles