A Study On The Influencing Factors And Forecasting Of The Performance Of Rewards Crowdfunding

In 2009,the first rewards crowdfunding platform was kicked in the United States.After that,the rewards crowdfunding platform developed rapidly in the world.It is gradually mature in Europe and the United States and extends to Asia,Central and South America,Africa and other regions.In July 2011,the first award-winning platform "named time" was established in China,which symbolized the beginning of our platform about crowdfunding.In the next few years,many different platforms about crowdfunding quickly were set up.According to incomplete statistics from mingjin website,as of the end of November 2015.In the amount of the project,the number of rewards crowdfunding is the largest.In this paper,3468 rewards crowdfunding projects in the public network(http://www.zhongchou.com)are used to analyze the influencing factors of rewards crowdfunding projects and forecast the performance of rewards crowdfunding projects.And it provides theoretical and empirical basis for the development of rewards crowdfundin in China.In this paper,the transaction data in the public network platform is used to analyze the influencing factors of the performance of rewards crowdfunding,and the performance of rewards crowdfunding is forecasted..The optimal scaling model can be used to analyze the significant influencing factors and non-significant influencing factors of rewards crowdfunding projects.And then the paper explores the reasons why these influencing factors are the significant influencing factors and non-significant influencing factors.In order to predict the financing performance of rewards crowdfunding projects,the SOM is used to discretize the financing performance.Then,the C5.0 decision tree,the support vector machine and the TAN Bayesian network are used to predict the financing performance.The C5.0 decision tree algorithm,support vector machine algorithm and TAN Bayesian network are compared and analyzed,and the prediction of C5.0 decision tree is the best.The results show that: it can be found that the length of the project name,the amount raised,the industry of the project,the city and the number of comments have effect on the financing performance.However,the remaining variables have less impact on financing performance.The C5.0 decision tree,the support vector machine and the TAN Bayesian network are used to predict the financing performance.Among the three algorithms,the C5.0 decision tree has the best predictive effect.In addition,except for Class 1 performance,the prediction results for C5.0 are higher than those of support vector machines and TAN Bayesian networks for predictions.Based on theoretical and practical research,for project financiers,for the rewards-crowdfunding project fundraisers,it is recommended that the name of the project should be named as reasonable and attractive.The target amount should be set according to the actual situation.The project initiator should promptly reply to the investor's comments and suggestions,with the aim of increasing the project initiator interacting with the investor.For the Rewards-crowdfunding project investors,we should pay attention to the development of the follow-up projects and the way of return about rewards crowdfunding.For the rewards-crowdfunding industry,it is recommended that the relevant departments of the regulatory need to further clarify the idea,the industry should gradually move toward the specification,rewarding platforms need to explore new models,rewards crowdfunding can be extended to some new industries.