DeepFINA. Deep Learning for Financial Investor Network AnalysisSkabelon

The inability of conventional financial models to meet empirical observations from real-world markets has led to a growing level of dissatisfaction. Existing models set unrealistic assumptions for investors’ behavioral characteristics and the information flow inside and outside the market.

The few recent interdisciplinary studies following data-driven analyses are based on crude behavioral investor categorizations, generic attributes for expressing properties of these investor categories (like age and gender), and simple (linear) mathematical models. The project will study financial activities at the individual investor level and its evolution over time, based on data-driven methodologies.

Principal Investigator from ECE, AU:

Henrik Karstoft

Professor (Docent)

M hka@ece.au.dk
H 5125, 312
P +4541893270

Grant source:
Danmarks Frie Forskningsfond, Teknologi og Produktion

Granted amount:
DKK 6.191.605

Project start:
01/09/2023

Revised 07.07.2026

Emma Peck-Thorsted