Analysis of Turkish textile sector with self organizing maps method: An application on the companies in Borsa İstanbul

Users of the financial statements need financial analysis to be able to make decisions on different issues such as credit, investment and management. The accessibility and complexity of financial data and limitations of analytical techniques prevent information users getting optimal benefit from the data they use.Statistical methods and artificial neural networks are used to overcome this situation and to increase the efficiency of financial analysis. In this study, it was aimed to show that the self-organizing maps algorithm, which is an unsupervised artificial neural network, is feasible and compatible with the financial analysis process. For this purpose, financial analysis was carried out by using annual financial statements and monthly stock prices between 2013-2018. Of the 16 corporations trading on the Istanbul Stock Exchange with corporations active in the textile sector. 15 financial ratios representing the liquidity, debt, activity, profitability and market valuation of the enterprises were used. The annual financial ratios and annual stock returns were presented to the algorithm as input data. The financial analysis of each year was done separately with the component planes and clusters obtained by the operation of the algorithm.