This demo uses symbolic regression via genetic programming to discover mathematical formulas that fit the target equation you define.
The algorithm evolves a population of candidate equations over multiple generations. Each generation, the best-performing equations are selected and combined to create new candidates. Over time, the population converges toward formulas that accurately represent the target function.
R² Score measures how well the prediction explains the variance in the data (1.0 = perfect fit). MAE shows the average absolute error between predicted and actual values.