基于投资者行为分析的众筹绩效预测模型

Crowdfunding performance prediction model based on investor behavior analysis

针对众筹融资过程中存在的信息不对称问题,基于前景理论中处理不确定信息的决策效用规则,结合众筹项目信息披露与投资者效用分析,构建了一个新的众筹绩效预测模型。为解决实际应用中特征选择过多的问题,引入了一种基于神经网络算法的稀疏性特征选择方法,该方法能够帮助众筹平台聚焦于核心特征,以更好地理解和预测投资者行为。对Kickstarter平台上超过15万个项目的实证分析结果表明:考虑投资者风险感知和前景效用的模型对众筹绩效有更好的预测和解释能力。该研究结论不仅为众筹项目的预测和评估提供了新的视角,也为众筹平台和筹资者建立分析投资者支持行为的分析框架提供了有力的工具。

Addressing the issue of information asymmetry in crowdfunding,this paper developed a new model for predicting crowdfunding performance,based on the decision utility rules for processing uncertain information in prospect theory and combining the analysis of crowdfunding project information disclosure with investor utility.To tackle the issue of excessive feature selection in practical applications,it introduced a sparsity-based feature selection method using neural networks,which could help crowdfunding platforms to focus on core features for better understanding and predicting investor behavior.Empirical analysis of over 150000 projects on the Kickstarter platform shows that models considering investors’perception of risk and prospect utility have better predictive and explanatory power for crowdfunding performance.The research results not only provide a new perspective for the prediction and evaluation of crowdfunding projects,but also offer powerful tools for crowdfunding platforms and fundraisers to establish models for analyzing backers’backing behavior.