Skip to main content

Table 1 Machine learning models and performance in nested cross-validation on training dataset – escape/ absconding vs. no escape/ absconding

From: Escape and absconding among offenders with schizophrenia spectrum disorder – an explorative analysis of characteristics

Statistical Procedure Balanced
Acc. (%) (95% CI)
AUC (%)
(95% CI)
Sensitivity (%) (95% CI) Specificity (%)
(95% CI)
PPV (%)
(95% CI)
NPV (%)
(95% CI)
Logistic Regression 73.9 [64.7, 79.8] 0.81 [0.72, 0.89] 80.2 [76.0, 84.4] 67.6 [52.5, 82.7] 95.8 [93.5, 98.1] 26.9 [17.9, 35.9]
Tree 72.6 [64.2, 79.6] 0.80 [0.71, 0.89] 81.9 [64.5, 92.3] 63.2 [47.8, 78.5] 95.3 [92.8, 97.7] 27.9 [18.4, 37.4]
Random Forest 71.1 [61, 76.6] 0.84 [0.76, 0.92] 94.8 [92.4, 97.1] 47.4 [31.5, 63.2] 94.2 [91.8, 96.7] 50.0 [33.7, 66.3]
Gradient Boosting 65.8 [58.4, 73.8] 0.82 [0.74, 0.90] 89.5 [86.3, 92.8] 42.1 [26.4, 57.8] 93.3 [90.6, 96.0] 30.8 [18.2, 43.3]
KNN 66.5 [58.6, 74.3] 0.85 [0.59, 0.99] 85.5 [81.7, 89.1] 47.4 [31.5, 63.2] 93.6 [90.9, 96.3] 26.5 [15.9, 36.9]
SVM 73.4 [65.5, 80.9] 0.84 [0.76, 0.92] 94.2 [91.7, 96.7] 52.6 [36.8, 68.5] 94.7 [92.3, 97.1] 50.0 [34.5, 65.5]
Naive Bayes 76.7 [68.4, 82.7] 0.88 [0.81, 0.95] 79.7 [75.4, 83.9] 73.7 [59.7, 87.7] 96.5 [94.3, 98.6] 28.6 [19.6, 37.5]
  1. AUC area under the curve (level of discrimination), PPV positive predictive value, NPV negative predictive value, KNN k-nearest neighbors, SVM support vector machines