This repository assists in the development and evaluation of deep learning models that use both Crypto Fear and Greed Index (FNG) values and simple closing prices to compare which one provides a better signal.
I used deep learning recurrent neural networks to model bitcoin closing prices. One model uses the FNG indicators to predict the closing price while the second model uses a window of closing prices to predict the closing price.
Steps completed:
- prepare-the-data-for-training-and-testing
- build-and-train-custom-lstm-rnns
- evaluate-the-performance-of-each-model
Which model has a lower loss? The LSTM Stock Predictor using closing prices has a lower loss of 0.0147 compared to 0.1065 for the FNG index.
Which model tracks the actual values better over time? The LSTM Stock Predictor using closing prices tracks the actual values better over time.
Which window size works best for the model? I found a window size of one predicts the price with the least amount of lag and lowest loss.
-
python
-
pandas
-
numpy
-
hvplot
-
tensorflow
-
sklearn
-
jupyter notebook