Symbolic regression

Symbolic regression searches the space of the mathematical functions for candidates for the optimal description of given data.

Genetic programming is often used to find the solution , where formulas are represented as expression trees, see picture.

Representation of a function as an expression tree . Sub-trees can be moved, changed or deleted (mutation) and complete trees can be combined (recombination).

Eureqa or HeuristicLab are examples of programs for symbolic regression.

Web links

https://towardsdatascience.com/symbolic-regression-and-genetic-programming-8aed39e7f030

Individual evidence

^ Ingrid Gerdes, Frank Klawonn, Rudolf Kruse: Evolutionary algorithms: Genetic algorithms - strategies and optimization methods - example applications . Springer-Verlag, 2013, ISBN 978-3-322-86839-8 ( google.de [accessed on July 20, 2020]).

[1] Ingrid Gerdes, Frank Klawonn, Rudolf Kruse: Evolutionary algorithms: Genetic algorithms - strategies and optimization methods - example applications . Springer-Verlag, 2013, ISBN 978-3-322-86839-8 ( google.de [accessed on July 20, 2020]).