Research Article

Evolutionary Framework with Bidirectional Long Short-Term Memory Network for Stock Price Prediction

Table 1

The structure of EBiLSTM and hyperparameter.

Input: real closing price (20 days)

[Layer1] internal cells (32) (BiLSTM | LSTM | GRU | RNN)
Output: predicted price in the next day
Hyperparameter: lr = 0.00001, epoch = 150, and timestep = 21
Survived parents number