Research On Intelligent Prediction Method Of Market Value Of Small And Medium-Sized Enterprises For Supply Chain Finance

Supply chain finance is a new financing model,which can alleviate the financing difficulties of small and medium-sized enterprises.However,risks in supply chain finance are prone to occur.In order to better control the risks,there are currently risk prediction methods using support vector machines and BP neural networks.However,the research on intelligent algorithms in the field of supply chain finance is not systematic enough.the performance of the algorithm needs to be improved.Therefore,this paper further improves the performance of the support vector machine and BP neural network by improving the particle swarm algorithm and the differential evolution algorithm.The main work of the paper is as follows:First,in view of the problem that the particle swarm algorithm cannot balance global search and timely convergence,this paper improves the inertia weight and acceleration factor of the particle swarm algorithm,and disturbs the particles in the later iteration of the algorithm.Firstly,this paper introduces the principle and existing problems of particle swarm optimization,and draws the conclusion that the acceleration factor and inertia weight should be increased or decreased respectively.Then,according to different optimization objects,this paper makes the particle swarm algorithm enhance the global search ability or speed up the convergence speed respectively.The method is to make the inertia weight and acceleration factor increase or decrease in the form of bump function respectively.Finally,experiments show that the proposed method has higher prediction accuracy.Second,this paper designs a new cooperative differential evolution algorithm based on immigration operator to solve the problem of unreasonable exchange of information between sub-populations in the multi-population differential evolution algorithm.First of all,this paper finds that the exchange of the number of immigration operators between subpopulations in the collaborative differential evolution algorithm is unreasonable,so a strategy of dynamic immigration operators is proposed.Then,the elite population is used for the evolution of sub-populations,and each sub-population adopts different evolution strategies to better balance the global search and local search capabilities.Finally,experiments show that the proposed method has higher prediction accuracy.Third,this paper implements a market value prediction system for small and medium-sized enterprises for supply chain finance.First of all,this paper makes a systematic analysis of the market value forecasting system,and designs the system according to the system analysis,including the design of the system technical architecture,the design of functional modules and the design of the database.Then,this paper implements each module of the system and deploys the above improved algorithm to the backend.Finally,the system is tested,and the test results show that each functional module can operate normally.