Skip to content

Latest commit

 

History

History
14 lines (12 loc) · 1008 Bytes

README.md

File metadata and controls

14 lines (12 loc) · 1008 Bytes

Exchange Rate Forecaster

Description

  • Daily foreign exchange data is pulled from a standard public REST API for the last 100 days.
  • The data contains daily open, high, low and close rates for the SEK-INR combination.
  • A machine learning model based on the LSTM neural network is used for training, validating and testing all the data available from 2014. The model is saved once and then used to make predictions given the latest 30 days data. The grey region shows the prediction for the next day.
  • Tensorflow is the main driver for bulding the LSTM model.
  • Checkout the jupyter notebook from model building/training.
  • All the tasks here are automated using GitHub Actions. See the ./github/workflows/automation.yml file.
  • Below is a Candlestick plot for the latest 30 days recorded and predicted values. alt forex monitor