A1888
Title: Forecasting realized volatility using machine learning: The case of EU energy listed firms
Authors: Cosmin Octavian Cepoi - The Bucharest University of Economic Studies (Romania) [presenting]
Alexandru Adrian Cramer - The Bucharest University of Economic Studies (Romania)
Roxana Clodnitchi - The Bucharest University of Economic Studies (Romania)
Daniel Traian Pele - Bucharest University of Economic Studies, Institute for Economic Forecasting, Romanian Academy (Romania)
Vasile Strat - The Bucharest University of Economic Studies (Romania)
Sorin Anagnoste - The Bucharest University of Economic Studies (Romania)
Abstract: The aim is to examine the effectiveness of AI techniques in forecasting realized volatility within a high-frequency data framework. Tick-by-tick data is utilized from 50 highly liquid companies in the energy sector, publicly traded on various European stock exchanges. The objective is to assess whether machine learning methods outperform traditional linear models in terms of predictive accuracy. Focusing on the latter half of 2021, a period marked by significant energy crises in Europe, the findings hold relevance for both industry practitioners and regulators. This analysis encompasses a period of increased volatility in European stock markets, offering valuable insights into the efficacy of AI for volatility forecasting during uncertain times.
