Abstract
Today, when national and international economic conditions are determinant in the determination of business activities, it is necessary for companies to use their assets and resources more effectively. The firms that do not want to face financial failure must identify their strengths and weaknesses by analyzing their performance in past periods before making strategic plans for their future activities. The aim of the study is to investigate and to determine the efficiency of 22 firms in BIST Food and Beverage Index by Data Envelopment Analysis (DEA) using efficiency scores of firms, and to determine the important in-firm variable or variables that effect the efficiency of firms through data mining techniques. For this aim 8 financial ratios are calculated by using the financial statements for the years 2013-2017. In the analyzes, 6 input variables, namely current ratio, acid-test rate, leverage rate, short term dept / total assets rate and fixed assets / total assets rate, long term dept / total assets rate; 2 output variables, total asset profitability and net profit margin, are used. By applying Data Envelopment Analysis efficient and inefficient firms are found. As a result of the study, 11 firms in 2014, 9 firms in 2015, 17 firms in 2016, 7 firms in 2017 and 11 firms in 2018 are found to be efficient. It is concluded that the most important variable affecting the effectiveness of the firms is net profit margin ratio according to Artificial Neural Networks and C5.0 Decision Tree technique.