Data mining application in prediction of financial failure for manufacturing businesses

Forecasting financial failure and related studies have been one of the areas that have attracted attention and gained wide coverage in the literature for more than forty years. Regarding the estimation of financial failure, which always maintains its importance both academically and sectorally; Many models have been developed using different data sets and methods. However, although there is no general estimation model that can be valid under all circumstances and conditions; Due to its importance, research on this subject continues at an increasing pace. In this study, it is aimed to develop a model that can predict the financial failure of businesses with a high percentage of accuracy and can be generalized for many situations. Between 2008 and 2017, the financial data of the production companies according to the transaction in the BIST (Borsa Istanbul) were used, and the methods of Decision Trees, Random Forests and Artificial Neural Networks, which are data mining techniques, were used to predict the failure of the relevant enterprises. In the study, besides classifying businesses as financially successful and financially unsuccessful; it was thought that it would be beneficial to consider the financially risky business class in terms of taking measures and developing strategies in line with the forecast. It is hoped that the study will differ in this aspect and contribute to the literature. In the study, six prediction models developed for the failure of the enterprises were differentiated according to the data reduction processes. In addition, the results obtained from the developed models were compared with the classical method.