Enhancing the Efficacy of Financial Information Through Artificial Intelligence

The use of artificial intelligence and machine learning techniques in finance is gaining more and more traction from practitioners as well as from academia. In fact, corporations nowadays are using these techniques to forecast and assess different financial risks such as liquidity risk, volatility risk, and credit risk by applying ML models. The ML models are trained on historical datasets to make future forecasts on potential financial threats to the financial performance of the company. Practitioners and institutional investors have been introducing artificial intelligence to assist their work and run different types of analysis based on quantitative and qualitative data. The introduction of qualitative (textual) data in financial market analysis is a relatively recent approach adopted by sophisticated investors to measure the tone, and the sentiment and extract information from corporate annual reports, press releases, and even social media posts. Natural Language Processing and text mining paired with machine learning models are still under trial but have proven to be effective in guiding sophisticated investors and corporate managers. Meanwhile, finance scholars were reluctant to introduce new methodologies, especially those relying on content and textual analysis for different reasons. Their orthodoxy not only in the way they write research but also in the topics they debate could be one of the reasons that probably makes their knowledge less accessible, sometimes less relevant, and probably not read by practitioners. Finance-related texts commonly meant to make information available to market participants, tend to be written in formal and technical language that makes them less intelligible than they should be, complicating the possibility to make sense of them and drive action in the financial environment for the great majority of individuals. This has, for a long time, been a silently accepted limit, though more recently it brought to attention the need for a more effective and transparent spread of financial information, aimed at reducing noise as a source of volatility. New technology, among them machine learning as a prominent application of Artificial Intelligence, maybe a handy instrument to underpin latent meanings, isolate prevalent emerging topics, and help non-professionals to make sense of financial information. Inaccessible information significantly reduces its real impact on the market’s dynamics, considerably limiting the possibility to enhance efficiency. In this work we introduce and describe in a very understandable way how machine learning may help improve the comprehension of financial information, we also present the results of our latest research, as a prominent example of how the application of machine learning to different fields may be of great utility both for the activity of scholars and researchers, but also for practitioners and investors.