主管:中国科学院
主办:中国优选法统筹法与经济数学研究会
   中国科学院科技战略咨询研究院

中国管理科学 ›› 2023, Vol. 31 ›› Issue (5): 71-83.doi: 10.16381/j.cnki.issn1003-207x.2020.1520

• 论文 • 上一篇    

奖励众筹融资绩效动态预测研究——来自“众筹网”数据的实证

张卫国1, 黄思颖2, 王超1   

  1. 1.华南理工大学工商管理学院,广东 广州510640;2.南方电网能源发展研究院,广东 广州510530
  • 收稿日期:2020-08-06 修回日期:2020-09-29 发布日期:2023-05-23
  • 通讯作者: 王超(1991-),男(汉族),陕西商洛人,华南理工大学工商管理学院,博士后,研究方向:金融工程与风险管理,Email:ccwang@scut.edu.cn. E-mail:ccwang@scut.edu.cn
  • 基金资助:
    国家自然科学基金-广东联合基金重点支持项目(U1901223);广东省自然科学基金研究团队项目(2017A030312001)

Research on Dynamic Forecasting of Rewarding Crowdfunding Financing Performance——Empirical Analyses from “Crowdfunding Network” Data

ZHANG Wei-guo1, HUANG Si-ying2, WANG Chao1   

  1. 1. School of Business Administration, South China University of Technology, Guangzhou 510640, China;2. Energy Development Research Institute of China Southern Power Grid, Guangzhou 510530, China
  • Received:2020-08-06 Revised:2020-09-29 Published:2023-05-23
  • Contact: 王超 E-mail:ccwang@scut.edu.cn

摘要: 奖励众筹融资绩效不仅受项目自身情况的影响,也与投资者行为有密切关系。本文研究了奖励众筹融资绩效与投资者投资和评论活动的动态变化关系,提出了投资曲线和评论曲线特征提取方法,并结合奖励众筹项目的投资特征、时滞评论特征以及项目自身特征,构建了奖励众筹项目融资绩效动态预测的函数主成分分析-广义回归神经网络(FPCA-GRNN)模型。基于“众筹网”交易数据进行了实证分析,结果表明FPCA-GRNN模型具有较好的融资绩效实时预测精度。

关键词: 奖励众筹;融资绩效;投资者行为;动态预测

Abstract: Starting in 2015, compared with foreign crowdfunding platforms that need to review the sponsor’s personal information, bank collection account, guarantor’s social security account, and other information, the Chinese platform has lower requirements for personal information disclosure of project sponsors, and the overall market has serious information asymmetry. Besides, most relevant studies are based on cross-sectional data, and few use panel data to analyze the dynamics of project financing. More importantly, existing research has not paid attention to the dynamic monitoring of rewarding crowdfunding financing performance and lacks corresponding measurement and monitoring methods. Limited by high information acquisition costs and low information processing capabilities, when faced with the virtuality and borderless nature of cyberspace, it is often difficult for investors to correctly evaluate project results based on existing project settings, description information, and real-time project financing progress which eventually generate the risk of investment failure. Based on this, an in-depth exploration of investor dynamic behavior is conducted, hoping to use innovative models to improve information processing capabilities and reduce the impact of information asymmetry in the crowdfunding market.

Key words: rewarding crowdfunding; financing performance; investors’ behavior; dynamic forecasting

中图分类号: