Research On Financial Risk Prediction Based On Hybrid Learning Strategy

With the enhanced trend of economic globalization, the market competition among enterprises is becoming more and more intensive, which increases the financial risk of enterprises. Financial risk not only damages the benefits of corporate investors and creditors, but also threatens the macroeconomic stability. Therefore, how to predict financial risk timely and effectively to help enterprises identify financial risk has become a problem needed to be solved imminently in academic and industry.At present, the researchers have done a lot of work on financial risk prediction, and put forward many financial risk prediction methods. However, it is difficult for these methods to obtain the ideal application effect. One of important reasons is that these methods don’t take data characteristics into full consideration. There are common problems such as unlabeled samples and imbalanced distribution of labeled samples in financial risk prediction applications. Therefore, this research combined semi-supervised classification and imbalanced data classification and proposed financial risk prediction model based on hybrid learning strategy to solve the above issues. Firstly, this research reviewed the research status of financial risk and machine learning systematically, and then clarified the research problems and future directions. Secondly, this research analyzed the basic theory of financial risk and hybrid learning, including the concept and characteristics of financial risk, the application of financial risk prediction, the concept of machine learning, the classification of machine learning and so on. Then, based on the above analysis, from the perspective of financial risk prediction application, this research applied financial risk prediction to corporate credit rating field and constructed corporate credit rating model based on hybrid learning strategy. Finally, this research applied it to financial distress prediction field. Financial risk prediction methods based on hybrid learning strategy were tested in these two specific applications to verify its effectiveness, experimental results indicates that hybrid learning strategy achieves the better results in the above two applications.In conclusion, on the one hand, the related theory in the field of enterprise financial risk prediction is analyzed systematically, and then the hybrid learning strategy have been proposed, which enriching the theory of enterprise financial risk prediction. On the other hand, this research applies the financial risk prediction model based on hybrid learning strategy to corporate credit rating and financial distress prediction field, which has provided an effective method for enterprise financial risk prediction.