LSTM Stock Predictor

Due to the volatility of cryptocurrency speculation, investors will often try to incorporate sentiment from social media and news articles to help guide their trading strategies. One such indicator is the Crypto Fear and Greed Index (FNG) which attempts to use a variety of data sources to produce a daily FNG value for cryptocurrency.

In this project I built and evaluated deep learning models using both the FNG values and simple closing prices to determine if the FNG indicator provides a better signal for cryptocurrencies than the normal closing price data.

One model will use the FNG indicators to predict the closing price while the second model will use a window of closing prices to predict the nth closing price.

Three Main Tasks:

1. Prepare the data for training and testing
2. Build and train custom LSTM RNNs
3. Evaluate the performance of each model

Prepare the data for training and testing

Use the starter code as a guide to create a Jupyter Notebook for each RNN. The starter code contains a function to create the window of time for the data in each dataset.

For the Fear and Greed model, you will use the FNG values to try and predict the closing price. A function is provided in the notebook to help with this.

For the closing price model, you will use previous closing prices to try and predict the next closing price. A function is provided in the notebook to help with this.

Each model will need to use 70% of the data for training and 30% of the data for testing.

Apply a MinMaxScaler to the X and y values to scale the data for the model.

Finally, reshape the X_train and X_test values to fit the model's requirement of samples, time steps, and features. (example: X_train = X_train.reshape((X_train.shape[0], X_train.shape[1], 1)))

Build and train custom LSTM RNNs

In each Jupyter Notebook, create the same custom LSTM RNN architecture. In one notebook, you will fit the data using the FNG values. In the second notebook, you will fit the data using only closing prices.

Use the same parameters and training steps for each model. This is necessary to compare each model accurately.

Evaluate the performance of each model

Finally, use the testing data to evaluate each model and compare the performance.

Use the above to answer the following:

Which model has a lower loss?

Closing prices model had a lower loss

Which model tracks the actual values better over time?

Closing prices model tracked actual values better over time

Which window size works best for the model?

Window size of 10 is the best size for these models.