The LSTM is an incredible ML architecture that is renowned for it's ability to learn long term dependences, making it extremely effective with time-series data. In this project I use historic S&P 500 stock data to train an LSTM to predict the future closing price of the stock for a given time.
- clean up spx_from_1950.csv, as it has a few null values.
- test my new network against the one built by Jakob Aungiers to assess efficacy of the extra features
This project builds on the on a youtube tutorial by Siraj Raval and some helper code by Jakob Aungiers who demonstrated how to create a model with one feature. I build on this by pre-processing more historical data and increasing the number of features from one to five.