Classifier | Models | Sensitivity | Specificity | f1_score | Precision | Accuracy | AUC |
---|---|---|---|---|---|---|---|
decision tree | Model 1 | 0.261(0.095) | 0.885(0.104) | 0.348(0.076) | 0.671(0.205) | 0.612(0.035) | 0.626(0.052) |
Model 2 | 0.448(0.145) | 0.817(0.082) | 0.514(0.147) | 0.638(0.191) | 0.664(0.052) | 0.628(0.077) | |
Model 3 | 0.340(0.081) | 0.805(0.103) | 0.417(0.097) | 0.570(0.171) | 0.606(0.077) | 0.570(0.077) | |
Model 4 | 0.412(0.127) | 0.826(0.077) | 0.489(0.137) | 0.632(0.186) | 0.655(0.049) | 0.615(0.071) | |
Model 5 | 0.487(0.186) | 0.776(0.108) | 0.525(0.175) | 0.608(0.206) | 0.658(0.072) | 0.658(0.104) | |
Model 6 | 0.410(0.209) | 0.799(0.104) | 0.451(0.192) | 0.618(0.213) | 0.63(0.079) | 0.615(0.081) | |
SVM | Model 1 | 0.265(0.129) | 0.932(0.048) | 0.403(0.080) | 0.738(0.123) | 0.639(0.080) | 0.547(0.049) |
Model 2 | 0.210(0.182) | 0.878(0.130) | 0.272(0.163) | 0.643(0.213) | 0.606(0.062) | 0.536(0.051) | |
Model 3 | 0.380(0.130) | 0.831(0.058) | 0.452(0.095) | 0.614(0.107) | 0.633(0.053) | 0.548(0.049) | |
Model 4 | 0.417(0.123) | 0.844(0.106) | 0.506(0.091) | 0.690(0.156) | 0.667(0.073) | 0.589(0.056) | |
Model 5 | 0.526(0.110) | 0.846(0.077) | 0.600(0.094) | 0.721(0.124) | 0.715(0.055) | 0.606(0.065) | |
Model 6 | 0.583(0.124) | 0.837(0.112) | 0.644(0.106) | 0.743(0.145) | 0.736(0.074) | 0.655(0.056) | |
random forest | Model 1 | 0.485(0.127) | 0.802(0.074) | 0.540(0.102) | 0.639(0.116) | 0.670(0.046) | 0.680(0.054) |
Model 2 | 0.443(0.118) | 0.838(0.106) | 0.523(0.110) | 0.690(0.190) | 0.673(0.048) | 0.735(0.099) | |
Model 3 | 0.425(0.121) | 0.772(0.095) | 0.474(0.082) | 0.579(0.123) | 0.621(0.051) | 0.679(0.045) | |
Model 4 | 0.414(0.130) | 0.883(0.089) | 0.511(0.118) | 0.740(0.164) | 0.688(0.027) | 0.769(0.073) | |
Model 5 | 0.504(0.152) | 0.866(0.111) | 0.587(0.154) | 0.743(0.218) | 0.712(0.095) | 0.780(0.074) | |
Model 6 | 0.514(0.109) | 0.869(0.098) | 0.597(0.105) | 0.756(0.174) | 0.718(0.062) | 0.779(0.075) | |
neural network | Model 1 | 0.343(0.102) | 0.876(0.090) | 0.441(0.097) | 0.688(0.178) | 0.648(0.064) | 0.676(0.060) |
Model 2 | 0.354(0.214) | 0.834(0.103) | 0.420(0.177) | 0.594(0.148) | 0.642(0.054) | 0.642(0.110) | |
Model 3 | 0.505(0.079) | 0.771(0.109) | 0.547(0.074) | 0.625(0.141) | 0.655(0.055) | 0.660(0.062) | |
Model 4 | 0.378(0.182) | 0.817(0.123) | 0.429(0.137) | 0.617(0.179) | 0.624(0.045) | 0.631(0.091) | |
Model 5 | 0.488(0.134) | 0.827(0.105) | 0.561(0.115) | 0.686(0.137) | 0.694(0.052) | 0.741(0.065) | |
Model 6 | 0.61(0.108) | 0.874(0.104) | 0.686(0.095) | 0.801(0.130) | 0.767(0.081) | 0.793(0.086) | |
naive bayes | Model 1 | 0.302(0.078) | 0.867(0.092) | 0.396(0.056) | 0.663(0.191) | 0.624(0.053) | 0.650(0.067) |
Model 2 | 0.279(0.134) | 0.868(0.101) | 0.371(0.132) | 0.634(0.172) | 0.624(0.087) | 0.613(0.096) | |
Model 3 | 0.480(0.119) | 0.858(0.068) | 0.562(0.087) | 0.718(0.096) | 0.700(0.042) | 0.767(0.064) | |
Model 4 | 0.279(0.134) | 0.868(0.101) | 0.371(0.132) | 0.634(0.172) | 0.624(0.087) | 0.647(0.093) | |
Model 5 | 0.385(0.101) | 0.907(0.065) | 0.503(0.091) | 0.770(0.096) | 0.688(0.070) | 0.761(0.078) | |
Model 6 | 0.385(0.101) | 0.907(0.065) | 0.503(0.091) | 0.77(0.096) | 0.688(0.070) | 0.771(0.072) | |
logistic regression | Model 1 | 0.351(0.098) | 0.838(0.104) | 0.433(0.079) | 0.630(0.167) | 0.627(0.054) | 0.667(0.065) |
Model 2 | 0.391(0.183) | 0.782(0.128) | 0.444(0.155) | 0.570(0.149) | 0.627(0.054) | 0.625(0.130) | |
Model 3 | 0.515(0.185) | 0.667(0.129) | 0.500(0.130) | 0.529(0.106) | 0.600(0.070) | 0.602(0.100) | |
Model 4 | 0.505(0.154) | 0.775(0.128) | 0.541(0.115) | 0.640(0.137) | 0.664(0.053) | 0.666(0.105) | |
Model 5 | 0.506(0.131) | 0.690(0.150) | 0.510(0.125) | 0.565(0.180) | 0.606(0.082) | 0.615(0.083) | |
Model 6 | 0.494(0.122) | 0.721(0.129) | 0.515(0.124) | 0.571(0.187) | 0.621(0.075) | 0.626(0.095) |