Research On Performance Evaluation Of Listed Companies Based On Optimized BP Neural Network

In the context of the country's strongly support for the e-commerce industry,in-depth study of B2 B e-commerce business performance evaluation,not only help to improve the level of e-commerce business performance evaluation,but also has a great guiding significance for the business managers,venture capitalists and the majority of shareholders to make a decision.Based on the understanding of the performance evaluation theory of domestic and foreign enterprises,this paper establishes the performance evaluation index system of B2 B e-commerce listed companies based on the characteristics of B2 B e-commerce listed companies,prepares the original data with factor analysis,and selects BP neural network as The performance evaluation method of this paper.However,BP neural network may be localized in the training process.In order to make up for this defect,this paper introduces the global search algorithm-genetic algorithm.On this basis,the GA-BP neural network model based on factor analysis is established,and the superiority of the model is verified by contrast experiment.Finally,the application of the model of domestic listed B2 B e-commerce listed companies to conduct a comprehensive evaluation.The main work of this paper is as follows:First of all,this paper studies the domestic and foreign literature on enterprise performance evaluation,and puts forward the performance evaluation index system which accord with B2 B e-commerce listed company,including financial and non-financial indicators.Secondly,this paper uses the factor analysis method to reduce the dimensionality of the original data,and uses the processed common factor as the input layer of BP neural network to achieve the purpose of simplifying the network structure and improving the efficiency of network training.The comprehensive score is used as the target output value of the training network to ensure the objectivity of the target value and avoid the influence of human factors.The BP neural network is used to determine the initial weights and thresholds randomly.The genetic algorithm is used to determine the optimal weights and thresholds of the BP neural network,so that the mapping function can reflect the actual situation more accurately,Based on this,a GA-BP neural network evaluation model based on factor analysis is established.Thirdly,The GA-BP neural network model based on factor analysis and GA-BP neural network model which is not analyzed by factor analysis are used to compare the experiment with GA optimized BP neural network model.The superiority of the evaluation model.Finally,Based on the index system established in this paper and the performance evaluation model,this paper makes a comprehensive evaluation of the performance of domestic listed B2 B e-commerce listed companies,and provides reference for relevant management,investors,fundraisers and the public.