Financial Risk Analysis and Early Warning Research Based on Data Mining and Statistical Analysis Technology

Abstract

Using the advantage of decision tree algorithm in the screening work, the traditional ID3 algorithm is improved and optimized, and a new and simplified financial index system is constructed. At the same time, combined with the unique value of artificial neural network in early warning model and data analysis, B-P model is used to build a mixed financial early warning model. In the model study, the HFPM model and Z-score model were compared and analyzed by using test samples and training samples, and the superior warning ability of the former was effectively verified.