Author(s)

Yihong Zhao

Corresponding Author

Yihong Zhao

ABSTRACT

With the improvement of computing power and the improvement of machine learning theory, the deep learning has a more significant advantage in its nonlinear fitting ability. Thanks to the mature methods of data processing, the deep learning can use a variety of algorithmic structures such as BP neural network and DB network to make use of the informative indicator on the network to improve the accuracy of the prediction results of financial risk management. The optimization method in the deep learning and the processing method of data noise can also promote the growth and improvement of relevant theories of financial risk management. However, the financial market is complex and changeable. Any models of deep learning will face unfavorable results or even opposite predictions. Therefore, it is very important to construct a model of deep learning with stable effects and high speed.

KEYWORDS

deep learning, back propagation network, deep brief network (DBN), convolutional neural network (CNN), financial risk management