A Model Based On Genetic Programming For Financial Crisis Forecast

In vigorous capital market, enterprise is able to collect the low cost fund from capital market and accelerate enterprise developing. The investor also can make use of capital market operation to invest, so he can gain higher reward. Modern enterprise is confronted with the gradually dangerous marketplace environment. Risk is perplexing enterprise director all the time in management. The survival process of modern enterprise is a process unceasingly producing and resolving various risks. It will affect the benefit to extensive investor considerably, once financial affairs crisis happened in enterprise. The purpose of the study is building a set of company financial affairs crisis forecast system.The main body of the study applies the genetic plan (GP) to financial affairs crisis forecast system. GP is a kind of evolution algorithm. It is based on genetic algorithm (GA). It is accomplished by Koza on 1992. It has overcome machine-made limitation of the particular tradition. It has suggested a kind of brand-new structure, whose essence is to use broad computer program to describe problem. This broad computer program is indicated the character strings structure together for tree shape in GP, and it changes its structure and size according to environment situation.The study quilts the annual Financial Crisis Company happened in 2000-2005 (ST company looks upon as financial affairs Crisis Company). The 100 financial affairs crisis companies have composed and the pairing's being not ST (The company dose not look upon as Financial Crisis Company). The 100 financial affairs crisis companies and the 106 being not ST companies are samples. The study has chosen Short-Term Sinking Ability, Working Ability, Long-Term Sinking Ability, Payoff Ability and Cash Flow. The five aspects include 20 financial targets variables in total. The study applies qualitative analysis method to choose the financial targets variable. The study uses thefinancial targets variable choosing to be an independent variable. The study uses the genetic plan method then. It uses the Discipulus software.Finally, the study comes to a conclusion, rightness rate of the genetic model is most high.