Business environments have been changed for years and more complicated than before. There are more impacts and many difficulties to companies because of financial distress. In order to reduce the impact, it is important to find out the causes of financial distress, and give alert to companies before it get distressed. There were studies that use basic financial analysis,
or single data-mining method exploring causes of financial distress. This study based on dynamic financial states, and finds the characteristics and rules of financial distress. This study combines Back-propagation Artificial Neural Networks (BPN) and Evolution Strategies (ES), extracts rules of financial distress, and incorporates with the results of Markov process analysis, in order to develop an optimized financial-distress alert model.
This study utilizes supervised learning networks to evaluate the risk of financial
distress of the companies, and find out the rules of financial distress by adding the prior evaluated results to evolution strategies.
Relation:
Proceedings of International Conference on Business and Information
