Empirical Study On Banking Security Warning Based On BP-neural Networks

Through several research on word-wide financial crises ,we found that if currency crises or financial crises occurred ,banking crises as flash point.When banking crises arosed concern of managers,it is too late,generally the bank have to face either closing down or taking over or insolvency. In our country ,commercial bank pay more attention to contraling the ongoing risk and ex post facto remediation, customarty ignoring risk management in advance.Researching and building up banking satey warning systerm ,momentarily abserving correlative index point,managing the possibility risk in advance ,which have signifacant meaning to avoiding the financial crises happening ,averting the negative effectiveness of financial crises erupting,furthermore it can ensure economy develepment in stability and healthy way.In this article,I expend research on banking safety from industrial perspective.Firstly thesis retrospects the study actuality of industrial safety and banking safety ,at the same time expatiates the research method and goal of warning theory; secondly thesis analise banking features on the industrial perspective,then expounds the factors which affect the banking safety from industrial economics,finally sets up a index systerm from three stages-macroeconomic,mesoeconomic combining qualitative method with quantitative method.thirdly thesis estates the banking safety status with principal components method based on 14 commercial bank which has get listed on the stock market;then on the basis of principal components result ,thesis inspects the result using back-propagation network, and warning the banking security; finally thesis brings forward suggestions on enhancing banking security. the empirical result shows,from 2004 to 2007,the record based on principal components method reproved year by year,at the same time passing the back-propagation network examination.