Evaluation of Back Propagation-Artificial Neural Network (BP-ANN) Fit Rate and Types of Vector Machine Algorithms in Estimating the Bankruptcy Prediction of Companies Listed on Tehran Stock Exchange

Main Article Content

Mehrdad Shafiee , Hossein Fakhari

Abstract

The accurate estimation of the bankruptcy prediction is an issue that has been increasingly considered regarding the importance of companies' bankruptcy prediction from investors, companies and banks (determining credit risk). However, determining new mathematical techniques that can provide a higher fit rate for this prediction requires further researches and comparisons of algorithms fit. Accordingly, the present study aimed at comparing the fit rate in vector machine algorithms and artificial neural network (ANN) to determine the companies' bankruptcy prediction in the coming years. The financial statements of companies over 2011-2019 (end of 2019) were reviewed to collect data. Article 141 of the Commercial Code was used to determine the bankrupt companies. Then, the desired algorithms were implemented by MATLAB software. The results showed that the support vector machine algorithms had a higher fit rate in estimating the companies' bankruptcy prediction (the maximum difference in model fit was 6%).

Article Details

Section
Articles