Table 2 Hyperparameters for the best models for predicting price direction DBMLA - Difference between Maximum and Lowest Accuracies
From: Bitcoin price change and trend prediction through twitter sentiment and data volume
LSTM | CNN | BiLSTM | |
|---|---|---|---|
# Layers | 1 | 3 | 2 |
Layer Size | 32 | 32 | 64 |
Batch Size | 80 | 50 | 80 |
Dataset | 1 day lag | 1 day lag | 1 day lag |
Lagged Features | 7 | 7 | 7 |
Train-Test Split | 85:15 | 85:15 | 85:15 |
Loss Function | Categorical Crossentropy | Categorical Crossentropy | Categorical Crossentropy |
Early Stopping Parameter | Validation Loss | Validation Loss | Validation Loss |
Early Stopping Patience | 20 | 20 | 20 |
Maximum Accuracy | 67.16% | 64.18% | 64.18% |
Mean Accuracy | 59.10% | 58.51% | 60.90% |
DBMLA | 14.93% | 11.94% | 7.46% |