moquan
TensorFlow LSTM Time Series Algorithm for Stock Market Forecasting
This inference model was trained to forecast the following 10 stocks on data from May 18th, 2012 to June 9th, 2023:
- AAPL
- AMD
- AMZN
- AVGO
- GOOG
- INTC
- META
- MSFT
- NVDA
- ORCL
Step 1:
Use the build_dataset script to convert individual ticker CSV data from Yahoo Finance's Historical Data tool to a complete training dataset.
Step 2:
Run the TensorFlow training algorithm, updating the ticker array and input shape as needed. By default, the input shape is (5, 10)
-- (10 input features and 5 previous time series point per prediction).
Dependencies (pip):
- tensorflow
- numpy
- pandas
- matplotlib
Step 3:
Update the ticker array in the model wrapper class as needed. Then, test predictions.