Product Crowdfunding Default Risk Warning Based on Random Forest Model

Abstract

At present, there are few researches about the default risk of product crowdfunding, and the default of product crowdfunding will cause great losses to the platform and investors. This paper studies the problem of default risk warning. The indicators used to construct the early warning index system of product crowdfunding default risk come from two parts, the characteristics of product crowdfunding and the commonality of product crowdfunding, P2P online lending, and equity crowdfunding. Finally, the default risk warning model of product crowdfunding was constructed based on the random forest. We find that the prediction accuracy of non-equilibrium data improves significantly with synthetic minority over-sampling technique. Besides, the sensitivity, specificity, accuracy, and AUC of the random forest model are higher than that of the support vector machine model. With the product crowdfunding default risk warning model, the product crowdfunding platform and investors can get better suggestions.