CREATES

A machine learning approach to volatility forecasting

Research output: Working paper/Preprint Working paper

Standard

A machine learning approach to volatility forecasting. / Christensen, Kim; Siggaard, Mathias Voldum; Veliyev, Bezirgen.

Aarhus : Institut for Økonomi, Aarhus Universitet, 2021.

Research output: Working paper/Preprint Working paper

Harvard

APA

Christensen, K., Siggaard, M. V., & Veliyev, B. (2021). A machine learning approach to volatility forecasting. Institut for Økonomi, Aarhus Universitet. CREATES Research Papers No. 2021-03

CBE

Christensen K, Siggaard MV, Veliyev B. 2021. A machine learning approach to volatility forecasting. Aarhus: Institut for Økonomi, Aarhus Universitet.

MLA

Christensen, Kim, Mathias Voldum Siggaard, and Bezirgen Veliyev A machine learning approach to volatility forecasting. Aarhus: Institut for Økonomi, Aarhus Universitet. (CREATES Research Papers; Journal number 2021-03). 2021., 47 p.

Vancouver

Christensen K, Siggaard MV, Veliyev B. A machine learning approach to volatility forecasting. Aarhus: Institut for Økonomi, Aarhus Universitet. 2021 Jan 18.

Author

Christensen, Kim ; Siggaard, Mathias Voldum ; Veliyev, Bezirgen. / A machine learning approach to volatility forecasting. Aarhus : Institut for Økonomi, Aarhus Universitet, 2021. (CREATES Research Papers; No. 2021-03).

Bibtex

@techreport{889c873d4d034aa0bc6f0f54864a8131,
title = "A machine learning approach to volatility forecasting",
abstract = "We show that machine learning (ML) algorithms improve one-day-ahead forecasts of realized variance from 29 Dow Jones Industrial Average index stocks over the sample period 2001 - 2017. We inspect several ML approaches: Regularization, tree-based algorithms, and neural networks. Off-the-shelf ML implementations beat the Heterogeneous AutoRegressive (HAR) model, even when the only predictors employed are the daily, weekly, and monthly lag of realized variance. Moreover, ML algorithms are capable of extracting substantial more information from additional predictors of volatility, including firm-specific characteristics and macroeconomic indicators, relative to an extended HAR model (HAR-X). ML automatically deciphers the often nonlinear relationship among the variables, allowing to identify key associations driving volatility. With accumulated local effect (ALE) plots we show there is a general agreement about the set of the most dominant predictors, but disagreement on their ranking. We investigate the robustness of ML when a large number of irrelevant variables, exhibiting serial correlation and conditional heteroscedasticity, are added to the information set. We document sustained forecasting improvements also in this setting.",
keywords = "Gradient boosting, High-frequency data, Machine learning, Neural network, Random forest, Realized variance, Regularization, Volatility forecasting",
author = "Kim Christensen and Siggaard, {Mathias Voldum} and Bezirgen Veliyev",
year = "2021",
month = jan,
day = "18",
language = "English",
series = "CREATES Research Papers",
publisher = "Institut for {\O}konomi, Aarhus Universitet",
number = "2021-03",
type = "WorkingPaper",
institution = "Institut for {\O}konomi, Aarhus Universitet",

}

RIS

TY - UNPB

T1 - A machine learning approach to volatility forecasting

AU - Christensen, Kim

AU - Siggaard, Mathias Voldum

AU - Veliyev, Bezirgen

PY - 2021/1/18

Y1 - 2021/1/18

N2 - We show that machine learning (ML) algorithms improve one-day-ahead forecasts of realized variance from 29 Dow Jones Industrial Average index stocks over the sample period 2001 - 2017. We inspect several ML approaches: Regularization, tree-based algorithms, and neural networks. Off-the-shelf ML implementations beat the Heterogeneous AutoRegressive (HAR) model, even when the only predictors employed are the daily, weekly, and monthly lag of realized variance. Moreover, ML algorithms are capable of extracting substantial more information from additional predictors of volatility, including firm-specific characteristics and macroeconomic indicators, relative to an extended HAR model (HAR-X). ML automatically deciphers the often nonlinear relationship among the variables, allowing to identify key associations driving volatility. With accumulated local effect (ALE) plots we show there is a general agreement about the set of the most dominant predictors, but disagreement on their ranking. We investigate the robustness of ML when a large number of irrelevant variables, exhibiting serial correlation and conditional heteroscedasticity, are added to the information set. We document sustained forecasting improvements also in this setting.

AB - We show that machine learning (ML) algorithms improve one-day-ahead forecasts of realized variance from 29 Dow Jones Industrial Average index stocks over the sample period 2001 - 2017. We inspect several ML approaches: Regularization, tree-based algorithms, and neural networks. Off-the-shelf ML implementations beat the Heterogeneous AutoRegressive (HAR) model, even when the only predictors employed are the daily, weekly, and monthly lag of realized variance. Moreover, ML algorithms are capable of extracting substantial more information from additional predictors of volatility, including firm-specific characteristics and macroeconomic indicators, relative to an extended HAR model (HAR-X). ML automatically deciphers the often nonlinear relationship among the variables, allowing to identify key associations driving volatility. With accumulated local effect (ALE) plots we show there is a general agreement about the set of the most dominant predictors, but disagreement on their ranking. We investigate the robustness of ML when a large number of irrelevant variables, exhibiting serial correlation and conditional heteroscedasticity, are added to the information set. We document sustained forecasting improvements also in this setting.

KW - Gradient boosting

KW - High-frequency data

KW - Machine learning

KW - Neural network

KW - Random forest

KW - Realized variance

KW - Regularization

KW - Volatility forecasting

M3 - Working paper

T3 - CREATES Research Papers

BT - A machine learning approach to volatility forecasting

PB - Institut for Økonomi, Aarhus Universitet

CY - Aarhus

ER -