A Neural Network Model to Predict Initial Return of Chinese SMEs Stock Market Initial Public Offerings

In this paper, a neural network model is developed to predict the first-day closing price of Chinese SME stock market IPOs. We try to predict the extraordinary Chinese IPOs first-day return by using a set of financial variables which is different from these existed research literatures considering the particularity of Chinese SME stock market issue mechanism and pricing method. We first pre-process the original data in order to eliminate the influence of different scale. Then we use principal component analysis to perform data reduction among all the post-processed data. Then we construct a neural network to predict the underpricing based on publicly available data in prospectus. The prediction result shows that it will be helpful to the investors´ decision making.