Design and Realization of Stock Forecasting Method Based on Long Short-term Memory Model

In order to accurately predict stock trends, the Long Short-Term Memory (LSTM) model was improved on the basis of Recurrent Neural Network (RNN), and three input gates, output gates, and forget gates were introduced. A control unit, the control unit in LSTM filters information when predicting data, so as to solve the long-term dependency problem of RNN. Experiments have shown that using LSTM technology and a stock prediction model designed and implemented based on the kaggle data set, the average error between the stock price prediction result and the actual price rise and fall is about 0.23, and the accuracy is high.