Abstract

Financial markets have become the most important part of every country's economic system, and the performance of financial markets will reflect the state of a country's economic development. In this paper, we use the gray system model and long and short-term memory network (LSTM) model to predict the stock price of GEM. Firstly, the paper introduces the algorithms related to gray system model and long and short-term memory network model, and proposes a new model hybrid scheme based on the LSTM model and GM(1,1) model with the core objective of stock price data prediction. This hybrid model can combine the advantages of the neural network model as well as the gray model and can get a better expected result around the stock price prediction problem. In terms of the fitting effect of the model, the fit of the gray system model is about 85%, the fit of the neural network model is about 86.3%, and the fit of the GM(1,1)-LSTM hybrid model is much improved and is above 90% overall. The GEM stock price prediction carried out in this paper is a very significant and valuable research problem in the current financial research field, and the effective analysis and prediction of stock prices can provide reference advice for investors and investment institutions in their investment decisions.