Description

TensorFlow implementation of the HARNet model for realized volatility forecasting.

Publication

R. Reisenhofer, X. Bayer, and N. Hautsch
HARNet: A Convolutional Neural Network for Realized Volatility Forecasting
arXiv preprint arXiv:2205.07719, 2022
https://doi.org/10.48550/arXiv.2205.07719

Please cite the paper above when using the HARNet package in your research.

Installation

Clone the repository and use

pip install -e HARNet/

to install the package.

Usage

Go to the HARNet root directory

cd HARNet

an start single experiments based on one of the preset configuration files

harnet ./configs/RV/RecField_20/HAR20_OLS.in
harnet ./configs/RV/RecField_20/QLIKE/HARNet20_QLIKE_OLS.in

Start experiments for all preset configuration files

/bin/bash run_all.sh

Results and TensorBoards for all experiments are saved in the ./HARNet/results folder.

About

The HARNet package was developed by Rafael Reisenhofer and Xandro Bayer.

The data in data/MAN_data.csv was obtained from the Oxford-Man Institute website.

If you have any questions, please contact [email protected].