LSTM-Luong algorithm based on entrepreneurial index closing price prediction research

Inhalt:
In order to study the future overall trend of newly created, high-growth enterprises in China, the trend prediction of the GEM index for its index price is made. In this paper, four algorithmic models, RNN model, LSTM model, LSTM_Bahdanau model, and LSTM_Luong model, are used to make predictions on the closing price data of the GEM index, aiming to find the most advantageous, more stable, more accurate prediction performance, and more suitable for the GEM index. Through comparison experiments, the LSTM_Luong model performs well in five indexes: mean absolute error (MAE), mean absolute percentage error (MAPE), root mean square error (RMSE), mean square error (MSE), and coefficient of determination, and the model is improved, with its MAE improving by 5.4 percentage points and RMSE improving by 3.5 percentage points, and its stability was improved.