Combining high frequency data with non-linear models for forecasting energy market volatility

作者：

Highlights：

• High-frequency data are coupled with nonlinear models to predict volatility.

• Comprehensive evaluation of multiple-step ahead volatility forecasts is provided.

• Neural networks provide both statistical as well as economic gains.

• Neural networks reduce tendency to over-predict volatility.

摘要

•High-frequency data are coupled with nonlinear models to predict volatility.•Comprehensive evaluation of multiple-step ahead volatility forecasts is provided.•Neural networks provide both statistical as well as economic gains.•Neural networks reduce tendency to over-predict volatility.

论文关键词：Artificial neural networks,Realized volatility,Multiple-step-ahead forecasts,Energy markets,C14,C53,G17

论文评审过程：Received 10 October 2015, Revised 17 January 2016, Accepted 7 February 2016, Available online 17 February 2016, Version of Record 4 March 2016.

论文官网地址：https://doi.org/10.1016/j.eswa.2016.02.008