Prediction On Stock Price Indexes Based On Grey Model And ANN

With the development of society and economy, stock has become one of the most important investment tools. As an investment tool with high risk and high return, the risk of stock is mainly considered as the volatility of its price. Stock price index reflects the over all volatility condition of the prices in the stock market, correct prediction on stock price index is helpful to make decision on stock investment.This paper studied the prediction of the stock price indexes of Shanghai and Shenzhen stock market based on ANN model and grey model.This paper sets up the prediction model of T+1, T+3 and T+5 based on the history data of day, week and month closing index. This paper calculated prediction precision index and direction correctness index of each model. In the process of setting up the models, several input data quantity were used to test the prediction performance, in order to study the relationship between quantity of input data and prediction precision level and direction correctness level.In this paper, the prediction precision levels of the models are very high, most of them below 10%, while the direction correctness levels are very low, most of them around 50%. Then, this paper studied on the factors which may affect the prediction precision level of the models. First, grouped the prediction precision indexes into several groups according to the factors; then, compared the relevant groups using nonparameter method to find out whether the factor has a significant influence on prediction precision. The result is that quantity of input data, prediction interval, data unit and index type all have a significant influence on prediction precision at a statistical level, while the factor of model type .does not have a significant influence on prediction precision at a statistical level.