A machine learning approach to volatility forecasting

Kim Christensen*, Mathias Voldum Siggaard, Bezirgen Veliyev

*Corresponding author af dette arbejde

Publikation: Bidrag til tidsskrift/Konferencebidrag i tidsskrift /Bidrag til avisTidsskriftartikelForskningpeer review

Abstract

We inspect how accurate machine learning (ML) is at forecasting realized variance of the Dow Jones Industrial Average index constituents. We compare several ML algorithms, including regularization, regression trees, and neural networks, to multiple Heterogeneous AutoRegressive (HAR) models. ML is implemented with minimal hyperparameter tuning. In spite of this, ML is competitive and beats the HAR lineage, even when the only predictors are the daily, weekly, and monthly lags of realized variance. The forecast gains are more pronounced at longer horizons. We attribute this to higher persistence in the ML models, which helps to approximate the long-memory of realized variance. ML also excels at locating incremental information about future volatility from additional predictors. Lastly, we propose a ML measure of variable importance based on accumulated local effects. This shows that while there is agreement about the most important predictors, there is disagreement on their ranking, helping to reconcile our results.
OriginalsprogEngelsk
Artikelnummernbac020
TidsskriftJournal of Financial Econometrics
Vol/bind21
Nummer5
Sider (fra-til)1680-1727
Antal sider48
ISSN1479-8409
DOI
StatusUdgivet - 2023

Fingeraftryk

Dyk ned i forskningsemnerne om 'A machine learning approach to volatility forecasting'. Sammen danner de et unikt fingeraftryk.

Citationsformater