Risk analysis with artificial neural network and regression in the finance sector

Banks are the most important element in financial markets. Banks have lots of important tasks with their functions. Banks both finance markets which demands funds and take funds from the suppliers. In addition to these banks should create funds to improve their processes. Banks work with liabilities so they take some of responsibilities so that risk management is very important for banks.Our thesis consists of four parts.In the first part, theoretical base of risk managament is evaluated and risk concept in finance sector is researched. Both Arragaments and standarts about the banking sector in Turkey and definition of risk management are explained in this study. In the second part, banks effected by risk groups which are studied on this work are researched and they are explained by theoretical frame. There are some technics that are necessity for observing, measuring and managament of risk groups.In the third part, Financial ratios demonstrate performance of the turkish banks which are active so that financial failure prediction models based on lojistik regression and artificial neural network model, wihich are among the multivariable statistical techniques, are developed for foreseeing financial failures. In the fourth part, there are comparisons and results of analysis that are obtained by the used models. Outputs about the predictions are evaluated in the last part of the work.Keywords: Financial risk, Financial failure, Financial performance ratios, Logistic regression, Artificial neural network, Statistical analysis.