Predicting Manager's Behaviour with BP Neural Network

There have been many researches on stock trend forecasting in existing literatures, however, there is little research on the actions which managers will take basing on stock price change. This article is based on this point to study the stock market and the actions that manager will take. Based on the proposed automatic selection model, we put it on the stock market to study whether the managers will sell shares in the company with the change of stock price. The results is that the selection of model has good generalization ability, and when the predicted value is higher than the actual value, the manager will sell off the company's shares, and vice versa. However, the accuracy of prediction will be further improved by combining factors such as company size and policy environment.