har-model

Julia package to generate, estimate, and forecast long memory processes

R code and Realized Volatility (RV) series set for fitting NN-based-HAR models to multinational RV series.

golang parsing and constructing HAR files

Tests whether option, news, attention and macro data beat HAR at forecasting SPY realized volatility, with LSTM, boosted trees and LASSO, Diebold-Mariano, MCS and Holm tests, then trades the forecasts in volatility targeting and delta-hedged straddles.

R package to estimate and forecast the HAR (Heterogeneous Autoregressive) model and its extensions.

Volatility-regime detection in equity index returns with a look-ahead-free Markov-switching model

Mini-Project Given during the Assignment 1 of ML Course of IITGN ES-335

Out-of-sample realized-volatility forecasting benchmark with HAR models, microstructure estimators, and Model Confidence Set testing.

Forecasting 2-hour FX realized volatility from 1-minute data with HAR-RV and HAR-RS models and walk-forward backtesting (EUR/USD 2012–2021)

Realized-volatility forecasting on 2y of Coinbase 5-min BTC/ETH: HAR/log-HAR/HAR-CJ vs GARCH-t vs EWMA, QLIKE, Diebold-Mariano, Model Confidence Set; log-HAR wins, GARCH 19-90% worse

Leakage-safe HAR/HARP realized-volatility forecasting with causal intraday-periodicity filtering and paired synthetic Monte Carlo evidence.

Forecasting BTC/ETH realized volatility from 4M one-minute bars: HAR, GARCH, GBM and LSTM under walk-forward validation with look-ahead tests, plus cost-aware volatility targeting.

Replication code and results for "Scale or speed? When do path signatures improve volatility-regime forecasts?" (Lyu & Lee)

Time-series econometrics project analysing IBM stock returns, ARMA-GARCH trading strategies, and HAR volatility forecasting using MATLAB.

Multi-agent LLM reading of financial news, measured against a HAR volatility baseline. Implementation of an M2 MIAGE ID dissertation (Universite Paris Dauphine-PSL).

Work in progress: can LightGBM beat HAR, GJR-GARCH, and EWMA at forecasting next-day SPY volatility?

Bi-LSTM + Attention vs HAR — comparative study of deep learning and econometric models in realized volatility forecasting (SPY, EWG, EPOL, EEM)

HAR-RS-DOW variance forecasting for BTC-EUR with Value-at-Risk, Expected Shortfall (Basel III), option pricing, and live trading deployment on Bitvavo.

⚡ Institutional NIFTY Realized Volatility Forecasting (HAR, Cluster/Sector-HAR, PCA, LightGBM, Diebold-Mariano) & Options Short-Strangle Liquidity Screener using DuckDB, FastAPI & Streamlit