Research On The Abnormal Changing Trend Of The Financial Status Of Listed Companies Based On Bayesian Network

The competed intense market-oriented economy, on the one hand provides a broad management stage for the companies, on the other hand again conceals innumerable rapid and dangerous shoal. Companies are very easy to be defeated if we do not give enough caution. Facing Chinese securities market, regardless of Chinese card supervises meets and Stock market which are regarded as the supervising and managing, or the broad investors it is all very important for taking a effective measure on the embryonic stage of financial status abnormity, to improve management and to prevent defeat if we can establish one set scientifically financial status change and trend of listed companies model on some method. As a result various benefit correlation can discovery some problems in financial management of listed companies promptly, realize the financial abnormal signal early, forecast financial status and trend of listed companies through the model.This paper can be divided into five parts. The first part is the introduction, which include the background and motivation of study, the domestic and international research of question survey. The second part introduced the basic principle of Bayesian Network. First of all, Bayesian statistics related knowledge are explained, then introduced elementary knowledge of Bayesian Network , Finally elaborates the characteristic of Bayesian Network analysis method. The third part explained study samples and how to decide financial analyzing variable. Through to limit the financial status abnormity of listed companies and the financial analysis variable selection of listed companies, determine the research samples of the paper, to analyze the financial status and future development trend of abnormal sample companies. The fourth part set up the Bayesian network, which was used for analyzing the abnormal change trend of financial status. Firstly, the sample companies of the financial status of the different development trend and the financial characteristic index are determined. Secondly, the node variable of Bayesian Network was indicated, and Bayesian Network chart was established. Thirdly the model of Bayesian Network was established. Finally the result was texted, based onthe Bayesian Network model of the financial status of the different development trend in the fourth chapter of the paper, inputs the financial index data of listed companies in 2004 respectively, to test the correct degree of the model. The fifth part is conclusion.Recently several approaches have been proposed for learning Bayesian Network. These methods combine prior knowledge with data to produce one or more Bayesian Network. The results of networks not only can be used for prediction, but also may infer causal relationships among variables. As a result, the resulting model can be applied in financial market to help analysts predict future market direction, understand critical market parameters, and make some investment decisions. By using Bayesian Network techniques this paper aims to discover some representative networks to analyze the financial situation change and development trend of listed companies from the viewpoint of fundamental analysis. Based on Bayesian network learning algorithms, the causes (financial ratios) and the effects (future development trend of listed companies) network representations are generated for encoding uncertain beliefs and drawing some conclusion from such representations. The empirical results show that the accuracy is about to 70%.