Forecasting the IBEX-35 Stock Index Using Deep Learning and News Emotions

Sergio Consoli,Marco Turchi,Elisa Tosetti,Amirhossein Tebbifakhr,Matteo Negri

doi:10.1007/978-3-030-95467-3_23

Sergio Consoli, Marco Turchi + Show 3 more

Open Access

https://doi.org/10.1007/978-3-030-95467-3_23

Copy DOI

Abstract

AbstractMeasuring the informational content of text in economic and financial news is useful for market participants to adjust their perception and expectations on the dynamics of financial markets. In this work, we adopt a neural machine translation and deep learning approach to extract the emotional content of economic and financial news from Spanish journals. To this end, we exploit a dataset of over 14 million articles published in Spanish newspapers over the period from 1st of July 1996 until 31st of December 2019. We then examine the role of these news-based emotions indicators in forecasting the Spanish IBEX-35 stock market index by using DeepAR, an advanced neural forecasting method based on auto-regressive Recurrent Neural Networks operating in a probabilistic setting. The aim is to evaluate if the combination of a richer information set including the emotional content of economic and financial news with state-of-the-art machine learning can help in such a challenging prediction task. The DeepAR model is trained by adopting a rolling-window approach and employed to produce point and density forecasts. Results look promising, showing an improvement in the IBEX-35 index fitting when the emotional variables are included in the model.

Highlights

Forecasting economic and financial variables is a challenging task for several reasons including, among others, the effect of volatility, regime changes, and low signal-to-noise ratio [15]
In this work we focus on the IBEX-35 index and Spanish news [14], the methodology is generalizable to other languages, and portable to other domains and evaluation scenarios
In this paper we present an approach aimed at exploring the predictive power of news for economic and financial time series forecasting

Summary

Introduction

Forecasting economic and financial variables is a challenging task for several reasons including, among others, the effect of volatility, regime changes, and low signal-to-noise ratio [15]. In this context, the incorporation in forecasting models of economic and financial information coming from news media, like in particular emotions and sentiment, has already demonstrated great potentials [1,4,8,23, 33, 42, 49]. Other papers ([46,47] and [41] among others) have classified articles in topics and extracted emotional signals that showed to have a predictive power for measures of economic activity, such as GDP, unemployment and inflation [25]. These results have shown the high potential of emotions extracted from news on monitoring and improving the forecasts of economic developments [17]

Methods

Results

Conclusion