Machine Learning Approaches for Intelligent Financial Decision Making

Financial decision making is a complex, multidimensional activity that aims at minimizing costs and maximizing returns, while remaining viable economically and socially. While significant advances have been made in other areas of study, this phenomenon presents itself as an underexplored field. Accordingly, and given the global pervasiveness of financial systems and their universal relevance, it becomes crucial to study financial decision making as an ongoing endeavor to ultimately delve into artificial intelligence. Machine learning models, including supervised, unsupervised, and reinforcement-learning algorithms-represent many of the most advance ideas in the modelling of financial decision-making systems, allowing informational structures to be identified and leveraged with respect to given goals. Adoption of these models in the financial sphere consequently warrants thorough investigation for the improvement of dissemination and practical advantages. Machine learning proves advantageous in financial decision making as well due to the openness and flexibility of its architecture, allowing for straightforward adaptation to heterogeneous, evolving situations. These aspects strongly relate to the dynamic nature of contemporary financial ecosystems: price trends, agent behaviors, systemic crises, regulation and governance, and the idiosyncrasies of individual firms can shift rapidly. Modeling therefore represents an ongoing endeavor and unpacking its fundamentals further fuels understanding of the financial decision-making phenomenon, thereby enhancing the incorporation of machine learning in financial contexts, encouraging transfers of principles and experiences from other sectors, and broadening the corpus of machine-learning-specific knowledge relevant to financial decision making.