Comparison of optimized machine learning approach to the understanding of medial tibial stress syndrome in male military personnel

Table 3 The performance of employed optimized machine learning approach in no resampling experiment based on the predictor’s importance and the number of predictors

# Predictors	Accuracy (%)	Sensitivity (%)	Specificity (%)	AUC	Selected ML Model
1	77.78	0	100	0.746	Naïve Bayes
2	85.19	50	95.24	0.766	Naïve Bayes
3	85.19	33.33	100	0.841	Naive Bayes
4	81.48	33.33	95.24	0.817	Naive Bayes
5	81.48	50	90.48	0.825	Naive Bayes
6	81.48	50	90.48	0.889	Naive Bayes
7	81.48	66.67	85.71	0.913	Naive Bayes
8	85.19	66.67	90.48	0.889	Ensemble
9	85.19	66.67	90.48	0.833	Naive Bayes
10	88.89	66.67	95.24	0.857	Naive Bayes
11	88.89	66.67	95.24	0.857	Naive Bayes
12	88.89	66.67	95.24	0.865	Naïve Bayes
13	85.19	50	95.24	0.802	Naïve Bayes
14	81.48	33.33	95.24	0.853	Naïve Bayes
15	85.19	50	95.24	0.786	Ensemble
16	85.19	50	95.24	0.802	Ensemble
17	85.19	50	95.24	0.849	Ensemble
18	85.19	50	95.24	0.825	Ensemble
19	81.48	33.33	95.24	0.825	Ensemble
20	85.19	33.33	100	0.929	Ensemble
21	81.48	33.33	95.24	0.786	Ensemble
22	85.19	50	95.24	0.873	Ensemble
23	85.19	50	95.24	0.889	Ensemble
24	81.48	33.33	95.24	0.825	Ensemble
25	81.48	33.33	95.24	0.929	Ensemble

ISSN: 1756-0500