Skip to content
/ moquan Public

TensorFlow LSTM Time Series Algorithm for Stock Market Forecasting.

Notifications You must be signed in to change notification settings

hershyz/moquan

Repository files navigation

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.


Training Results (view in notebook):

AAPL AMD AMZN AVGO GOOG INTC META MSFT NVDA ORCl

About

TensorFlow LSTM Time Series Algorithm for Stock Market Forecasting.

Resources

Stars

Watchers

Forks

Packages

No packages published

Languages