Econo- and Sociophysics Seminar

Finance and economics lack the experimental capabilities we know from physics. As a result, it is difficult to evaluate different models and theorems, which concern financial markets. One of the solutions for this problem is the agent based modelling (ABM), which was proposed already in the 20th century, but gained more momentum at the beginning of the 21st century, with the development of the computation power and algorithmic capabilities. The idea is simple, let's model financial entities and market participants as individual agents that live in a synthetic, virtual environment, similar to the real market. These agents were based on simple, but reasonable rules and in some cases were able to reproduce complex macroscopic behaviours we know from real-world markets. But can people or institutions be modelled by a few simple rules and assumptions?In my talk I will focus on modelling the limit order book and stock market in general. I will comment on recent developments and interesting observations, which confirm the big potential behind ABM in finance. Finally, I will show some promising directions, especially the ones related to modern machine learning tools, which may help to answer the question from the previous paragraph. https://us02web.zoom.us/j/82025295160?pwd=Rno0Skx6Q3haRzVzZy90dkRrbmxQQT09 Meeting ID: 820 2529 5160 Passcode: 794269