I. Almalis, E. Kouloumpris, and I. Vlahavas. “Sector-level sentiment analysis with deep learning”. Knowledge-Based Systems, p. 109954. issn: 0950-7051., 2022, DOI: https://doi.org/10.1016/j.knosys.2022.109954.
This paper presents new machine learning methods in the context of Natural Language Processing (NLP), in order to extract useful information from
financial news. Traditional NLP approaches, based on the use of lexicons or
standard machine learning algorithms, have ignored the importance of position and word combination in texts, resulting in reduced performance. More
recently, NLP empowered by deep learning has achieved remarkable results in
various tasks, such as sentiment analysis. This paper proposes a deep learning solution for sentiment analysis, trained exclusively on financial news, that
combines multiple recurrent neural networks. Following this, our sentiment
analysis models are further enhanced via a semi-supervised learning method
that relies on the detection and correction of presumably mislabeled data.
The performance of our proposed solution is compared favourably against
both traditional and state-of-the-art models based on its performance of previously unseen tweets data. The paper also provides novel research towards
the prediction of the specific economic sectors affected by news articles. Finally, we propose an ensemble of the sentiment and sector models in order
to provide sector-level sentiment analysis with potential applications in the
context of sector fund indices.
