Astrophysics Source Code Library

Making codes discoverable since 1999

ASCL Code Record

[ascl:2310.011] AI-Feynman: Symbolic regression algorithm

AI-Feynman fits analytical expressions to data sets via symbolic regression, mapping the target variable to different features supplied in the data array. Using a neural network with constraints in the number of parameters utilized, the code provides the ability to obtain analytical expressions for normalized features that are used to predict a Pareto-optimal target. AI-Feynman is robust in handling noisy data, recursively generating multidimensional symbolic expressions that match data from an unknown functions.

Code site:
Used in:
Described in:

Views: 841

Add this shield to your page
Copy the above HTML to add this shield to your code's website.