Symbolic Regression in Financial Economics

Jiacheng Liu, Siqi Guo

Keywords: symbolic regression, dataset, benchmark, deep learning, genetic programming

TL;DR: We present a new symbolic regression database consisting of equations from financial economics.

Abstract: We apply symbolic regression, the machine learning approach of recovering models from data, in financial economics. Specifically, we present a data set consisting of equations that cover a broad range of topics in financial economics. These equations are built off a common set of mathematical symbols but importantly have new variations in functional forms. We test the joint performance of deep learning and genetic programming symbolic regression systems in recovering these non-physical equations.