Research On The Prediction Model Of Capital Based On Wavelet Neural Network

With the development of global economic integration and the increase of economic uncertainty in the future,the traditional industry boom analysis is no longer applicable to the current complex forms,while the demand for capital flow of enterprises increases significantly.This year,novel coronavirus pneumonia triggered global systemic risks,which made people pay more attention to the grey rhinoceros and black swan event.Therefore,it is based on the continuous development of the reality and the changes of potential risks,and the demand for future cash flow forecast by enterprises and individuals is increasing.The higher this is,this article is based on Alipay data flow of funds as the research object,exploratory analysis of capital forecasts.In this paper,the basic mathematical principles are combined to get wavelet neural network.The advantages and combination principles of wavelet neural network are explained in detail.Secondly,the time series modeling of wavelet neural network is carried out by two-part analysis method to get linear separated wavelet and non-linear embedded wavelet neural network.Finally,this paper analyzes the fund purchase index of 28000 users,uses the above model to predict the Shanghai Stock Exchange 50 information and the user flow information of yu'ebao,constructs the prediction information result of the model,and then uses the Shanghai Stock Exchange 50 result on the prediction day as the input of the new model to predict the cash flow.It is concluded that the performance of embedded wavelet neural network is better.In this paper,the classical model support vector machine model and the wolf swarm optimization algorithm model are added to make a comparative analysis of the fund prediction results,which proves the feasibility and prediction accuracy of the innovative model in this paper.The implementation of this model can provide services for capital forecasting enterprises,so it has certain significance.