Designing a business risk forecasting model using machine learning techniques

One way to help you capitalize on investment opportunities and better allocate resources is to anticipate business risk. Predicting the probability of future events based on present and past information, In this way, first of all, by providing the necessary warnings, companies can be alerted to the occurrence of business failure so that they can take appropriate action accordingly. And second, investors and lenders distinguish favorable investment opportunities from unfavorable ones. And invest their resources in the right opportunities; Therefore, predicting the business risk of companies has always been one of the topics of concern for investors, creditors and the government. The purpose of this study is to design a business risk forecasting model using machine learning techniques.. The statistical population of the present study is the selected companies listed on the Tehran Stock Exchange during a period of nine years between 2007-2018. Hypothesis test results show that firm size, firm liquidity, firm profitability, firm growth opportunity, industry size, number of firms in industry have a negative effect on business risk. Meanwhile, the company's debt ratio has a positive and significant effect on business risk. The results also show that statistically, the life of the company, the decentralization of the industry has no effect on business risk. The results of model design and machine learning techniques show the efficiency of NB technique and then SVM technique compared to other machine learning techniques.