The Study On Financial Early-Warning For Listed Manufacturing Companies Based On BP Neural Network Optimized By Genetic Algorithm

With the level of development of market economy to improve, the market is morecompetitive. Companies should face more crises and challenges. For a large number of listedcompanies in the maufacturing sector, the financial issues are particularly important. If thesecompanies fall in financial trouble, it will not only threaten their own survival anddevelopment, but also bring significant losses to investors, creditors and numerousstakeholders.Therefore, the early-warning research and exploration of the financial positionhas important significance.A large number of domestic and foreign literatures have been able to prove that thefinancial indicators can play a certain role in indicating the enterprise’s financial situation. Butthe financial early-warning index system is imperfect. There are also many deficiencies andlimitations in so many financial early-warning models. This paper is aimed at buliding theearly-warning model and early-warning index system.This paper summed up28early-warning indexes which have significant impact on thefinancial position by analying the result of previous studies. It contained financial indexes andnon-financial corporate governance indexe. To build up the index system provided the basis inselection of financial early-warnig model. This paper selected the BP Neural Network modelwhich has wide range of application in financial early-warning in intelligent algorithms. In theempirical study, we used the Genetic Algorithm to optimize BP Neural Nerwork in order tocompensate the drawback that weights and thresholds are easy to fall into local minimum inthe revision process. It could make the early-warning model more efficacious.The empirical analysis is divided into two steps. The first step is to do early-warninganalysis with the BP Neural Network. The second step is to do forecast with the optimized BPNeural Network by the Genetic Algorithm. By comparing the results of the two-step, the BPNeural Network optimized by the Genetic Algorithm can have the more significant results. Itimproves the prediction accuracy and shortens the modeling time. This new method proposesa new thinking of financial warning in the theory and practice study.