Applying the Deep Learning Algorithm in Order to Provide a Model for Predicting the Financial Risk Management of Companies with a Quantitative Approach

Abstract

Companies face various risks during their business and financial cycle. These risks can be divided into different categories so that they can be identified and evaluated more easily. Therefore, it is important to monitor and accept changes caused by structural failures in the risk management process. The purpose of this research is to use deep learning algorithm in order to provide a model for predicting the financial risk of companies. Therefore, it is practical in terms of purpose. In terms of information gathering method, the research was library based and based on literature and theoretical background. It was also a quantitative research approach. Therefore, there is a need to provide a community and local model of risk forecasting that fits the financial and economic structure of companies active in Iran's capital market. In a small part of the statistical community, there were companies active in the capital market of Iran.The statistical sample was based on the method of systematic targeting of 199 active companies in the stock market between 2013 and 2014. The results of this research can lead to the expansion of the theoretical foundations of past researches related to predicting the financial risk of companies admitted to the stock exchange in Iran. In addition, the evidence of the research will show that this issue as a scientific achievement can provide useful information to the compilers of financial standards as well as the supervisory institutions of the country. Also, the results of the research can suggest new ideas for conducting new researches in the field of financial engineering, financial management and economics. The results showed that the error values of the training models in the deep learning approach in all cases of Lasso regression, ridge regression, artificial neural network training. And the random forest regression was less than 0.05.And the best method for machine learning is to use the combined method of ridge regression and artificial neural network.