Last phonon dos peak - cm-1 - last_phdos_peak#

ARFS Top features#

ARFS selected descriptors#

Relevant Descriptors


Correlation analysis#

Distance correlation#

Distance correlation heatmap

Dependency graphs#

Dependency graph

Feature learnability#

Feature learnability


Model performance#

5-Fold CV Metrics overview#

RF - MATMINER

train_rmse

test_rmse

train_errors

test_errors

train_r2

test_r2

mean

32.7937

84.3866

14.6748

39.5086

0.9956

0.96906

min

30.4141

55.9272

14.2786

31.0462

0.9951

0.954

max

34.5876

106.683

14.9702

44.2828

0.9961

0.9853

std

1.47065

18.1411

0.251989

4.91545

0.00032249

0.0127694

RF - MATMINER+LOBSTER

train_rmse

test_rmse

train_errors

test_errors

train_r2

test_r2

mean

32.5001

80.4644

14.6361

39.1976

0.99564

0.97226

min

26.9725

61.1432

13.7767

33.5999

0.9949

0.9492

max

35.1927

115.35

15.3407

45.2728

0.997

0.9824

std

2.91981

18.3066

0.512474

4.11005

0.000728286

0.012148

MODNet - MATMINER

train_rmse

test_rmse

train_errors

test_errors

train_r2

test_r2

mean

20.4586

67.3629

11.1928

33.3479

0.99828

0.97842

min

18.6992

37.2775

10.5934

24.3813

0.998

0.9416

max

22.5584

123.586

11.9387

46.2481

0.9985

0.9935

std

1.38787

29.3642

0.475595

7.19291

0.000203961

0.018805

MODNet - MATMINER+LOBSTER

train_rmse

test_rmse

train_errors

test_errors

train_r2

test_r2

mean

22.174

62.5857

10.7228

30.6388

0.99792

0.982

min

17.8546

43.5458

9.5286

27.3127

0.9972

0.9533

max

26.3936

110.553

12.5559

36.6207

0.9986

0.9911

std

3.52212

24.3505

1.23771

3.4089

0.000577581

0.0143975

Corrected resampled t-test on 10-fold CV#

Summary

t_stat

p_value

significance_stars

d_av

rel_improvement

percent_folds_improved

RF

0.123344

0.452273

0.0150349

0.317817

50

MODNet

1.02389

0.166309

0.26788

5.40627

70

RF t-test

MODNet t-test


Model Explainer#

PFI#

RF pfi MODNet pfi

SHAP#

RF shap MODNet Shap


Misc#

ARFS n-iter convergence checks#

Convergence

MAE/ fold from 10-fold CV#

Alternative visual summary of input data for t-test

RF per fold MAEs

MODNet per fold MAEs