Browsing by Author "Ural, Kerem"

Now showing 1 - 1 of 1

İflas tahmin yöntemlerinin karşılaştırılması: Borsa İstanbul üretim şirketlerinde uygulama
(2020) Ural, Kerem; Yeşilova, Fatma Dilvin Taşkın
Since Beaver's first studies of estimating bankruptcy ,studies on the estimating of bankruptcy of enterprises have been an attention grabbing field for many economists and scientists. Early detection of bankruptcy and financial difficulties has become the most critical part of financial analysis. From investors and enterprises to the government and persons, bankruptcy and financial difficulties have many negative influences. This situation has been seen clearly especially in the rise of exchange rates in 2018. It has been a cause for showing more importance on the estimation studies of bankruptcy. However, when we look at the present studies, we see that every model consists of distinctive variables. As these studies change both from sector to sector and nation to nation, they have been limited only with academic studies. In order to prevent this situation and ease the use of it for the sector, the studies between 1960-2018 have been analysed. The variables used in these studies have been grouped under seven factors. Five variables have been chosen for each factor. For further analysis, the number of variables have been decreased to one. In the study, the failure estimation model has been formed with seven variables. The main purpose of the study is, instead of a model only used in academic studies, to form one that can be used by the sector for making failure estimations with low cost and little effort. To ensure this, we performed a study by using the data of the manufacturing companies operand in the Istanbul stock exchange (BIST) between 2014 and 2018 and made a 5 years early bankruptcy estimation. In this study, we used three different learning models- one theoretical, one statistical and one machine. As theoretical model, we made the estimation of balance adverse; as the statistical model, we made logistic regression and as machine learning we made estimation with artificial neural network. The results of these models have been compared with each other. It has been understood that the bankruptcy expenditures of the enterprises are as important as the early estimation of the bankruptcy. In the study, we also tried to calculate the indirect expenditures that occur during the financial difficulty stage. To reveal these, we determined the expenditures of bankrupt enterprises between 2012 and 2014. We searched the market and profit lost together with the increase in the financial expenses of the bankrupt companies.