Table 5 Mean Square Error (MSE) and Mean Absolute Error (MAE) between actual and predicted responses using 3-fold cross validation for different category classification approaches (A, B, C mentioned in details in Description) for different drugs of CCLE database.

From: Heterogeneity Aware Random Forest for Drug Sensitivity Prediction

Drug Name

Cancer types

Number of samples

 

RF

A. HARF

B. Prior Classification of trees

C. LDA

AZD-6244

CNS & Skin

29 & 40

MSE

0.0164

0.0150

0.0153

0.0155

MAE

0.1038

0.0927

0.1002

0.0937

Lapatinib

Skin & Breast

40 & 29

MSE

0.0092

0.0078

0.0095

0.0086

MAE

0.0742

0.0648

0.0750

0.0689

Nilotinib

HLT & LUNG

69 & 74

MSE

0.0185

0.0177

0.0189

0.0180

MAE

0.0849

0.0831

0.0850

0.0841

PD-0325901

CNS & Skin

29 & 40

MSE

0.0259

0.0212

0.0250

0.0252

MAE

0.1340

0.1083

0.1297

0.1193

Panobinostat

CNS & HLT

29 & 71

MSE

0.0082

0.0073

0.0079

0.0073

MAE

0.0752

0.0686

0.0737

0.0686

Topotecan

HLT & Skin

71 & 40

MSE

0.0159

0.0150

0.0164

0.0174

MAE

0.1009

0.1001

0.1033

0.0152

  1. (T = 100, m = 10 and n size  = 4).