Abstract
Using a moderately large number of financial ratios, we tried to build models for classifying the companies listed on the Bucharest Stock Exchange into low and high risk classes of financial distress. We considered four classification techniques: Support Vector Machines, Decision Trees, Bayesian logistic regression and Fisher linear classifier, out of which the first two proved to have the highest prediction accuracy. Classifiers were trained and tested on randomly drown samples and on four different databases built starting from the initial financial indicators. As the literature related to the topic on Romanian data is very scarce, our study, by using a variety of methods and combining feature selection and principal components analysis, brings a new approach to predicting financial distress for Romanian companies.
Highlights
Companies’ performance is a widely spread research subject and is of interest for researchers and practitioners alike
Using a database of over 6900 companies, 28 financial ratios and 3 country specific indicators along with methods like Altman’s Z-Score, features selection and support vector machines, authors conclude that macroeconomic variables are relevant for bankruptcy prediction, considering that the accuracy rates are high for most models
Even though our approach to the problem of financial distress is different than that of other researchers, as we considered the high risk group as containing firms which either were in financial distress in 2015 or faced the problems in the subsequent two years, the results show high accuracy rates for all classifiers
Summary
Companies’ performance is a widely spread research subject and is of interest for researchers and practitioners alike. The methodologies approached in this respect range from very simple indicator-like scores and go to sophisticated supervised learning data mining techniques and neural networks Most studies in this area of research are focused on identifying a pattern in financial data which can predict future distress for companies by considering indicators with one to five years prior to the bankruptcy date, the financial indicators are calculated taking into account the bankruptcy date for companies. Accuracies over 85% in average for test datasets (30% of total companies) show that the financial indicators we used provide quality information about he risk of distress This could be a starting point for the development of a general model that may be applied on Romanian companies from different markets. The paper is structured as follows: literature review, methodological aspects, data and preliminary analysis, results and discussions and conclusions
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
