Abstract:

Purpose: The purpose of this study is the use of machine learning algorithms to evaluate the continuity of businesses. For this purpose, the data derived from the 136 companies constantly listed in Borsa Istanbul between the years 2010-2019 are used. Companies whose data could not be accessed or with different taxonomy were not included in the study. Approach: Artificial neural networks, decision tree, support vector machines, random forest, k-nearest neighbor classification, logistic regression and gaussian naive bayes algorithms were used in the study. The artificial neural networks and support vector machines used in the study work as black boxes. Other algorithms used in the study are rule-based. Class balanced 10-fold cross validation method was used in the application of the methods. Findings: As a result of the analysis, the overall success rates of decision tree and random forest algorithms were determined as 91.2% and 91.1%, Type 1 error 7.1% and 7.6%, Type 2 error 13.2% and 12.2%. In addition, return on assets ratio, ratio of retained earnings to total assets, financial leverage ratio, ratio of cash flow amount to total liability and current ratio variables were determined as important variables to evaluate the continuity of businesses. Originality: Numerous methods have been used in the literature to evaluate the continuity of businesses. However, in recent years, machine learning has come to the fore. In Turkey, the number of studies conducted with machine algorithms in the evaluation of the continuity of businesses is very few. In this study, the most used algorithms were applied together. Thus, the most successful algorithms were determined.

Keywords: