Learning Curve

import random

from compas_surrogate.surrogate.diagnostics.learning_curve import plot_learning_curve_for_lnl

OUTDIR = "out_learning_curve"
H5 = "det_matrix.h5"
random.seed(1)

if __name__ == "__main__":
    plot_learning_curve_for_lnl(
        OUTDIR, H5, 0,
        [100, 300, 500, 700, 1000]
    )

|COMPAS-SUR|24/08 12:18:43|INFO| Generating cache out_learning_curve/cache_lnl.npz using det_matrix.h5 and universe 0:MockPopulation(<Universe: [11735 systems], SF=0.006,2.770,2.900,4.233, muz=-2.66e-01, sigma0=2.73e-01>) 
|COMPAS-SUR|24/08 12:18:43|INFO| Starting LnL computation for 10667 universes 

|COMPAS-SUR|24/08 12:18:57|INFO| Keeping 9921/10667 likelihoods (dropping nans/inf). 
|COMPAS-SUR|24/08 12:18:57|SUCCESS| Saved out_learning_curve/cache_lnl.npz 

[learning_curve] Training set sizes: [ 100  300  500  700 1000]

[Parallel(n_jobs=2)]: Using backend LokyBackend with 2 concurrent workers.

Learing curve plot

Plot of “Mock” universe used for Lnl computation