Skip to main content

Table 4 The detailed predictions on the All Tumor dataset. All Tumor dataset: testing result of the Disc-GS-KNN classifier by setting 160 gene clusters to select 80 genes. 36 out of the 46 predictions were correct (in bold).

From: Selecting dissimilar genes for multi-class classification, an application in cancer subtyping

# Samples BR PR LU C LY BL ML U LE R PA O MS CNS
Breast (BR) 3 2      1         
Prostate (PR) 2   0     1         
Lung (LU) 3    2        1     
Colorectal (C) 3     3           
Lymphoma (LY) 6      6          
Bladder (BL) 3       2 1        
Melanoma (ML) 2        2        
Uterus-adeno (U) 2         2       
Leukemia (LE) 6          5 1     
Renal (R) 3    1    1     1   1   
Pancreas (PA) 3     1        2    
Ovary (O) 3       1       2   
Mesothelioma (MS) 3              3  
CNS 4               4