Skip to main content

Table 5 Comparison of prediction error on data with noise for the different models

From: Supervised learning methods in modeling of CD4+ T cell heterogeneity

Noise level

Approach

IL17

RORgt

IFNγ

Tbet

FOXP3

Sum of prediction error

Uniformly distributed noise in range of [−0.5 %, 0.5 %]

Artificial Neural Network

0.0671

0.0698

0.042

0.0362

0.0354

0.250

Linear Regression

0.235

0.235

0.190

0.129

0.0355

0.824

Support Vector Machine

0.0329

0.146

0.182

0.178

0.111

0.649

Random Forest

0.0413

0.0479

0.0364

0.0769

0.0397

0.242

Uniformly distributed noise in range of [−1 %, 1 %]

Artificial Neural Network

0.0706

0.0553

0.0435

0.0361

0.0393

0.2448

Linear Regression

0.795

0.682

0.677

0.546

0.46

3.16

Support Vector Machine

0.179

0.177

0.147

0.112

0.0406

0.6556

Random Forest

0.0552

0.0495

0.0484

0.0935

0.0349

0.2815