Table 3 Cross-validation (CV) performance summary.

From: VENturing into machine learning for the morphological analysis of von Economo neurons

Method

2-Fold CV

5-Fold CV

Average standard deviation

CNN methods

 Original

92.8% ± 1.2%

92.4% ± 1.5%

1.35%

 Diameter-enhanced

96.2% ± 0.8%

96.8% ± 1.2%

1.00%

 Soma-focused

94.1% ± 1.1%

94.3% ± 1.4%

1.25%

Traditional ML

 Random Forest

94.6% ± 1.8%

94.2% ± 2.1%

1.95%

 XGBoost

94.1% ± 2.0%

93.8% ± 2.3%

2.15%

 SVM

92.4% ± 2.5%

92.1% ± 2.8%

2.65%

  1. Results show mean accuracy ± standard deviation across folds. Lower average standard deviation indicates more stable performance.