Artificial Neural Network Models for Pricing Initial Public Offerings
利用公开财务数据构建人工神经网络模型来定价首次公开募股，发现该模型能带来显著的经济收益，并通过多项测试验证了结果的普适性和稳健性。
In recent times, managerial applications of neural networks, especially in the area of financial services, has received considerable attention. In this paper, neural network models are developed for a new application: the pricing of Initial Public Offerings (IPOs). Previous empirical studies provide consistent evidence of considerable inefficiency in the pricing of new issues. Neural network models using publicly available financial data as inputs are developed to price IPOs. The pricing performance and the economic benefits of the neural network models are evaluated. Significant economic gains are documented with neural networks. Several tests to establish generalizability and robustness of the results are conducted.