Generating alpha with AI in the stock markets : Towards understanding the process
Hulkkonen, Janne (2019-08-06)
Generating alpha with AI in the stock markets : Towards understanding the process
(06.08.2019)
Julkaisu on tekijänoikeussäännösten alainen. Teosta voi lukea ja tulostaa henkilökohtaista käyttöä varten. Käyttö kaupallisiin tarkoituksiin on kielletty.
suljettu
Julkaisun pysyvä osoite on:
https://urn.fi/URN:NBN:fi-fe2019081924627
https://urn.fi/URN:NBN:fi-fe2019081924627
Tiivistelmä
Artificial intelligence (AI) is a high-profile topic – it is currently impossible to avoid in the media. It is said that a technological arms race exists in the field of finance, regarding AI. The subject of this thesis has been researched relatively little with a qualitative approach. Therefore, this thesis was written with a qualitative approach to explore the phenomenon of AI in stock investing more broadly than usually seen in the literature, from the perspectives of experts on the subject. The purpose of this thesis is to provide clarity to the process of generating alpha with AI in the stock markets and also to
the main aspects to take into account when trying to or considering to use AI in stock investing effectively.
The empirical research methods used in this study have been semi-structured interviews and content analysis. Four experts were interviewed, and five different online contents (text, audio, and video) based on the views of experts on the subject were analyzed. The most recurring and relevant themes regarding the process of generating alpha with AI in the stock markets were (abbreviated): recruiting suitable, high skilled personnel, choosing the suitable stock investing and AI techniques, acquiring and refining suitable data, choosing a suitable computing strategy, and after the project has started, developing the AI stock investing solution. The main aspects to take into account when trying to, or considering to use AI in stock investing effectively, were found to be black boxes and how to treat them, how to execute backtesting and avoid overfitting, considering how autonomous an AI system should be, and if the considered business can survive in the near future without AI’s help.
It seems that the practical results of AI-based stock investing have not been strong, probably because the practice appears to still be in its beginning. However, the sector is growing actively, research results are promising, and some speculate that the financial markets might change because of AI. Thus, researchers and practitioners should familiarize themselves with the subject.
the main aspects to take into account when trying to or considering to use AI in stock investing effectively.
The empirical research methods used in this study have been semi-structured interviews and content analysis. Four experts were interviewed, and five different online contents (text, audio, and video) based on the views of experts on the subject were analyzed. The most recurring and relevant themes regarding the process of generating alpha with AI in the stock markets were (abbreviated): recruiting suitable, high skilled personnel, choosing the suitable stock investing and AI techniques, acquiring and refining suitable data, choosing a suitable computing strategy, and after the project has started, developing the AI stock investing solution. The main aspects to take into account when trying to, or considering to use AI in stock investing effectively, were found to be black boxes and how to treat them, how to execute backtesting and avoid overfitting, considering how autonomous an AI system should be, and if the considered business can survive in the near future without AI’s help.
It seems that the practical results of AI-based stock investing have not been strong, probably because the practice appears to still be in its beginning. However, the sector is growing actively, research results are promising, and some speculate that the financial markets might change because of AI. Thus, researchers and practitioners should familiarize themselves with the subject.