Research On Influencing Factors And Dynamic Prediction Methods Of Reward-Based Crowdfunding Financing Performance

Reward-based crowdfunding is an innovative financing method developed from the traditional financing disintermediation.It has the advantages of various forms,flexible investment,low cost,convenient transaction and wide coverage.The financing through the Internet can effectively reduce the financing cost and expand the scope of financing,which reduces the development restrictions of SMEs to the extent of “funding difficulties and financing”.However,with the deep reshuffle of the crowdfunding industry in 2017,China's reward-based crowdfunding faces a problem of low completion rate and a significant reduction in investor participation compared to last year.Therefore,how to optimize and improve the project,increase the participation of investors,and thus improve the financing performance is of great significance.It can enable crowdfunding to better serve the various economic segments of the real economy,and realize the crowdfunding financial model service assets and funds at both ends.This paper mainly studies from the following aspects:Firstly,based on the ELM model,this paper establishes a multivariate regression model to analyze the impact factors of crowdfunding financing performance from three aspects: central path and edge path,project quality signal,investor participation and uncertainty measurement.factor.Results show that the performance of rewarding crowdfunding financing is related to the target financing amount,the number of praises,the number of comments,the number of supporters,the number of updates,and the type of reward.Secondly,this paper constructs a fixed effect model using panel data to analyze the dynamic evolution process of rewarding crowdfunding financing performance.The results found that the early financing performance of project financing declined rapidly,and by the end of project financing,with the impact of the cutoff effect,the new financing performance showed a slight increasing trend;indicating that the investment behavior of the leading investors has a squeeze on the investment behavior of potential investors.The effect is that the higher the realized financing performance,the weaker the investor's new investment behavior.Finally,from the perspective of dynamic change,this paper constructs a comprehensive scoring model based on functional principal component analysis(FPCA)on the characteristics of investor investment activities and the characteristics of review activities considering the influence of time lag,and verifies the financing projects through correlation test.The existence of a crowdfunding success project has a strong investment and a similarity in the characteristics of the comments considering the effects of time delay.Furthermore,the comprehensive feature of the investment characteristics of the crowdfunding project,the comprehensive feature of the comment feature with the influence of time delay and the project characteristics are used as input variables to construct a generalized regression neural network(GRNN)model to realize the dynamic prediction of the financing performance of the crowdfunding project.Finally,the application research of “Crowdfunding Network” data shows that the FPCA-GRNN model constructed in this paper can realize real-time prediction of project performance,even in the case of small amount of data in the early stage of crowdfunding project,it can achieve better prediction accuracy.