Skip to main content

Table 3 Performance results for SI

From: Phenotype prediction from genome-wide association studies: application to smoking behaviors

SI

 

Prediction method

  

Feature selection method

# of SNP

LR

SVM

RF

EN

LDA

LR

100

0.7597

0.7171

0.7067

0.7670

0.7605

 

500

0.8792

0.9139

0.8132

0.9038

0.8964

SVM

100

0.6819

0.6421

0.6204

0.6794

0.6813

 

500

0.7953

0.8178

0.6930

0.7943

0.8075

RF

100

0.5961

0.6101

0.5980

0.5848

0.5957

 

500

0.6185

0.6312

0.6138

0.6010

0.6210

EN

100a

0.7930

0.7708

0.7336

0.7929

0.7 937

 

163a

0.8157

0.8084

0.7454

0.8188

0.8180

LDA

100

0.6338

0.5925

0.5807

0.6273

0.6343

 

500

0.7387

0.7503

0.6212

0.7176

0.7464

  1. In each column, the best results are shown as underlined. In each row, the best results are boldfaced. a. # of SNPs with non-zero coefficient.