Table 3 Overview of NMJ diagnosed using machine learning.
From: NMJ-Analyser identifies subtle early changes in mouse models of neuromuscular disease
Tenfold RF | Morphological features | Type | Overall permutation | Overall Gini | |
|---|---|---|---|---|---|
Accuracy | 0.9552 | Shape factor of endplate | Shape | 100 | 47.2 |
95% CI | (0.9417, 0.9664) | Shape factor of nerve | Shape | 58.39 | 18.6 |
P value [Acc > NIR] | < 2 × 10−16 | Intersection | Interaction | 35.1 | 100 |
Kappa | 0.8604 | Internal rugosity of nerves | Shape | 34.1 | 31.3 |
Mcnemar's Test (P value) | 0.6774 | Internal rugosity of endplates | Shape | 31.5 | 17.7 |
Sensitivity | 0.8809 | External rugosity of nerves | Shape | 27.4 | 22.88 |
Specificity | 0.9741 | Coverage | Interaction | 24.8 | 40.1 |
Prevalence | 0.2024 | External rugosity of nerves | Shape | 24.1 | 24.7 |
Balanced Accuracy | 0.9275 | Nerve non-compactness | Shape | 23.3 | 43.2 |
Surface/volume ratio of nerves | Size | 21.1 | 24.2 | ||
