Correlation versus RMSE Loss Functions in Symbolic Regression Tasks
Neural and Evolutionary Computing
2022-08-02 v2
Abstract
The use of correlation as a fitness function is explored in symbolic regression tasks and the performance is compared against the typical RMSE fitness function. Using correlation with an alignment step to conclude the evolution led to significant performance gains over RMSE as a fitness function. Using correlation as a fitness function led to solutions being found in fewer generations compared to RMSE, as well it was found that fewer data points were needed in the training set to discover the correct equations. The Feynman Symbolic Regression Benchmark as well as several other old and recent GP benchmark problems were used to evaluate performance.
Cite
@article{arxiv.2205.15990,
title = {Correlation versus RMSE Loss Functions in Symbolic Regression Tasks},
author = {Nathan Haut and Wolfgang Banzhaf and Bill Punch},
journal= {arXiv preprint arXiv:2205.15990},
year = {2022}
}
Comments
Submitted to the GPTP conference