Skip to main content

Table 2 Total prediction accuracy (TPA), true positive rate (TPR) and true negative rate (TNR) with kNN, Lasso, SVML and SVMR in Brown Swiss and Fleckvieh cattle

From: Use of SNP genotypes to identify carriers of harmful recessive mutations in cattle populations

  Method TPA TPR TNR
Brown KNN 0.837 (±0.047) 0.862 (±0.062) 0.798 (±0.088)
  Lasso 0.988 (±0.010) 1.000 (±0.000) 0.968 (±0.029)
  SVML 0.991 (±0.010) 0.998 (±0.005) 0.976 (±0.025)
  SVMR 0.978 (±0.016) 0.989 (±0.015) 0.961 (±0.034)
  MAG 0.983 (±0.015) 0.992 (±0.011) 0.984 (±0.016)
Fleckvieh KNN 0.970 (±0.007) 0.288 (±0.087) 0.998 (±0.002)
  Lasso 0.998 (±0.002) 0.957 (±0.037) 0.999 (±0.001)
  SVML 0.996 (±0.002) 0.924 (±0.047) 0.999 (±0.001)
  SVMR 0.991 (±0.003) 0.775 (±0.073) 0.999 (±0.001)
  MAG 0.994 (±0.003) 0.991 (±0.003) 0.998 (±0.002)
  1. KNN k-nearest neighbors, Lasso 1-penalized logistic regression, SVML suport vector machine with linear kernel, SVMR support vector machine with radial kernel MAG multi-allelic gene prediction