Analyzes and forecasts Microsoft's stock prices using Python and LSTM neural networks on Yahoo! Finance’s massive dataset
- Processed Microsoft's stock price data since 1986 taken from Yahoo! Finance's dataset of Microsoft with pandas, creating functions to define windows of dates for the purpose of splitting the dataset
- Divided the dataset into training, validation, and testing, with a split of 80/10/10, in that respective order
- Added the validation set for the purpose of tuning hyperparameters in the neural network
- The neural network included 1 Long Short Term Memory (LSTM) layer, 2 hidden Dense layers, and 1 node as output. Trained for 100 epochs
- Below, you can see the similarity in prediction and observation stock prices through the graphs
