The Study Of Financial Distress Prediction For The Listed Companies

Under the market economy condition with intense competition, as a result of various reasons, the financial risk is inevitable. Even some scale listed companies also face the problem of financial risk. Because the financial situation of the listed companies relates the benefits of all aspects, how to measure the financial risk accurately, and provide against it has become the focal point of some aspects such as: the bank, the investor as well as the government supervisory department.Foreign theoretical papers in this regard starts relatively earlier in the study, and has already made notable achievements in this area. But we are at an early stage in the regard, and the theoretical models are not very mature, the research system mostly uses the foreign study results for reference.This paper conducts the fundamental research of financial distress, and selects some listed companies in our country as the samples to conduct the empirical study and the contrastive study. The author has done the following several aspects of the work.Firstly, carries on the summary of the research condition in this domain both in the domestic and abroad, introduces the domestic and foreign. research condition in detail and carries on the appraisal to it. Upon this, proposes the research technique and the train of thought. Secondly, introduces the basic theories of Principal Components Analysis and the BP neural network, and selects 106 stocks from Shanghai and Shenzhen stock market as the research samples according to some principles and methods, divides them into the training group and the examination group, then, establishes the target system according to some methods. Finally, establishes the prediction model combining the Principal Components Analysis method and the BP neural network. After the empirical study and the comparative study, we can get that: the distinction accuracy of the combing model is 85.71%, which is higher than the sole BP neural network model 64.29%. From this also obtains that the Principal Components Analysis and the BP neural network union model is more suitable to make the financial distress prediction, simultaneously also explains that the listed companies in our country has acceptable financial index in quality. In the end of the paper, proposes some factors which may influence the findings, and makes a forecast to following research.