Globally Optimal Symbolic Regression
This provides a globally optimal solution for symbolic regression, which could benefit researchers in fields like physics and data science, though it appears incremental as it builds on existing optimization methods.
The authors tackled the problem of symbolic regression by introducing a technique that guarantees global optimality, formulating it as a mixed integer non-linear program to find minimum complexity expressions, and demonstrated it by rediscovering Kepler's law and Galileo's pendulum equation with real data.
In this study we introduce a new technique for symbolic regression that guarantees global optimality. This is achieved by formulating a mixed integer non-linear program (MINLP) whose solution is a symbolic mathematical expression of minimum complexity that explains the observations. We demonstrate our approach by rediscovering Kepler's law on planetary motion using exoplanet data and Galileo's pendulum periodicity equation using experimental data.