Financial Market Crash and Phase Transition: Through Model-Free Framework with Machine Learning

Abstract

Recent studies have attempted to understand market crash using the concept of phase transition in statistical physics. This study finds certain behavior in the financial market such as the critical phenomena that occur during phase transition. We apply model-free framework using convolutional neural network methods instead of the complex mathematical models studied previously. The results show that the financial market crash has a similar behavior to the phase transition of particles. Furthermore, we find that the similar behavior between financial market crash and phase transition gives better understanding on the market crash and detecting it.