In this project, we sought to develop Machine Learning (ML) models that predict future forex (FX) movements and sentiment of news headlines. The aim of the project was to create a one-stop-shop FX platform for investors and businesses to get the latest FX rates and news, as well as bidirectional price signals and news sentiment from our ML models to make better decisions on when to make Forex transactions.
.
├── data
│ ├── raw
│ ├── intermediate
│ ├── processed
│ └── temp
├── results
│ ├── outputs
│ ├── models
│ └── weights
├── documents
│ └── images
├── notebooks <- notebooks for explorations / prototyping
│ ├── news
│ └── signal
└── src <- all source code, internal org as needed
- Model weights:
Model weights are placed in the
weightsfolder
-
Clone this repo as follows
git clone <THIS_REPO_SSH/HTTPS>
-
Create the virutal environment
conda create -n openbank python=3.7.11
-
Activate the virutal environment
conda activate openbank
-
Install the requirements by running
python3 -m pip install -r requirements.txt
This project was deployed on a website to show the Bidirectional Forex Signals from the LSTM model and the News Sentiment of Financial News Headlines with FinBERT. Please access the website here