Table 6 Bootstrap validation results. Global classification metrics obtained for models A and B. Median accuracies (ACC) and F1-scores (F1) are presented with respective 95% CI. Best performing model on each scheme appears in bold
From: Machine learning based study for the classification of Type 2 diabetes mellitus subtypes
Models A | Models B | ||||
|---|---|---|---|---|---|
Scheme | Algorithm | ACC (95% CI) | F1 (95% CI) | ACC (95% CI) | F1 (95% CI) |
S1 | SVM | 0.9862 (0.978–0.993) | 0.9862 (0.978–0.973) | 0.9794 (0.969-0.987) | 0.9794 (0.969-0.987) |
KNN | 0.9292 (0.912–0.947) | 0.9284 (0.910–0.946) | 0.9307 (0.910-0.947) | 0.9298 (0.909-0.947) | |
MLP | 0.9880 (0.981–0.994) | 0.9880 (0.981–0.994) | 0.9832 (0.972-0.991) | 0.9832 (0.971-0.991) | |
SNNN | 0.9462 (0.928–0.962) | 0.9463 (0.928-0.962) | 0.8177 (0.782-0.845) | 0.8053 (0.759-0.836) | |
S2 | SVM | 0.8271 (0.807–0.846) | 0.8232 (0.802–0.843) | 0.9835 (0.974-0.990) | 0.9835 (0.975-0.990) |
KNN | 0.8074 (0.785–0.828) | 0.8023 (0.778–0.824) | 0.9542 (0.938-0.968) | 0.9539 (0.937-0.968) | |
MLP | 0.8166 (0.795–0.836) | 0.8154 (0.793–0.835) | 0.9891 (0.979-0.995) | 0.9891 (0.979-0.995) | |
SNNN | 0.8131 (0.788–0.837) | 0.8106 (0.782–0.836) | 0.8037 (0.766-0.830) | 0.7871 (0.729-0.819) | |
S3 | SVM | 0.7801 (0.759–0.803) | 0.7762 (0.753–0.800) | 0.8927 (0.876-0.908) | 0.8908 (0.874-0.907) |
KNN | 0.7735 (0.752–0.797) | 0.7687 (0.746–0.794) | 0.8751 (0.856-0.893) | 0.8735 (0.853-0.891) | |
MLP | 0.7643 (0.742–0.786) | 0.7625 (0.740–0.784) | 0.8934 (0.876-0.909) | 0.8917 (0.874-0.908) | |
SNNN | 0.7777 (0.750–0.802) | 0.7760 (0.744–0.801) | 0.7446 (0.705-0.774) | 0.7287 (0.659-0.763) | |
S4 | SVM | 0.9613 (0.949–0.972) | 0.9611 (0.949–0.972) | 0.9788 (0.969-0.987) | 0.9788 (0.969-0.987) |
KNN | 0.8921 (0.872–0.911) | 0.8902 (0.870–0.910) | 0.9037 (0.882-0.924) | 0.9023 (0.881-0.924) | |
MLP | 0.9781 (0.968–0.986) | 0.9781 (0.968–0.986) | 0.9833 (0.971-0.991) | 0.9833 (0.971-0.991) | |
SNNN | 0.9041 (0.882–0.922) | 0.9044 (0.882–0.922) | 0.7770 (0.736-0.808) | 0.7563 (0.689-0.792) | |
S5 | SVM | 0.8222 (0.803–0.842) | 0.8177 (0.797–0.839) | 0.9772 (0.968-0.985) | 0.9772 (0.968-0.985) |
KNN | 0.7948 (0.773–0.818) | 0.7863 (0.760–0.811) | 0.8988 (0.877-0.919) | 0.8974 (0.875-0.919) | |
MLP | 0.8149 (0.793–0.835) | 0.8140 (0.792–0.834) | 0.9851 (0.974-0.992) | 0.9851 (0.974-0.992) | |
SNNN | 0.8049 (0.778–0.831) | 0.8011 (0.769–0.828) | 0.7740 (0.736-0.805) | 0.7535 (0.684-0.791) | |
S6 | SVM | 0.7937 (0.763–0.819) | 0.7808 (0.742–0.811) | 0.9806 (0.971-0.988) | 0.9806 (0.971-0.988) |
KNN | 0.7553 (0.731–0.778) | 0.7310 (0.698–0.759) | 0.9504 (0.934-0.965) | 0.9502 (0.933-0.964) | |
MLP | 0.8190 (0.797–0.838) | 0.8175 (0.796–0.837) | 0.9810 (0.970-0.989) | 0.9810 (0.970-0.989) | |
SNNN | 0.7704 (0.742–0.795) | 0.7608 (0.727–0.789) | 0.9145 (0.848-0.956) | 0.9127 (0.841-0.955) | |
S7 | SVM | 0.7579 (0.737–0.781) | 0.7379 (0.709–0.766) | 0.9771 (0.968-0.985) | 0.9771 (0.968-0.985) |
KNN | 0.7431 (0.720–0.766) | 0.7216 (0.691–0.749) | 0.9312 (0.914-0.948) | 0.9306 (0.913-0.947) | |
MLP | 0.7398 (0.713–0.763) | 0.7296 (0.705–0.754) | 0.9785 (0.966-0.987) | 0.9785 (0.965-0.987) | |
SNNN | 0.7455 (0.717–0.772) | 0.7348 (0.693–0.763) | 0.8062 (0.765-0.839) | 0.7944 (0.727-0.831) | |