A comprehensive model for assessing and classifying patients with thrombotic microangiopathy: the TMA-INSIGHT score

Table 2 Performance metrics of machine learning model in the test set to predict thrombotic microangiopathy (TMA) in hospitalized patients

.metric	.estimator	XgBoost	LightGBM	Lasso
accuracy	multiclass	0.9778157	0.9795222	0.9385666
kappa	multiclass	0.9679064	0.9703983	0.9124146
Sensibility	macro	0.8223741	0.8575002	0.7572324
specificity	macro	0.9962477	0.9965055	0.9911797
ppv	macro	0.9120769	0.9340179	0.7446433
npv	macro	0.9969382	0.9971120	0.9897623
mcc	multiclass	0.9680054	0.9704651	0.9127704
j_index	macro	0.8186219	0.8540056	0.7484121
bal_accuracy	macro	0.9093109	0.9270028	0.8742060
detection_prevalence	macro	0.1250000	0.1250000	0.1250000
precision	macro	0.9120769	0.9340179	0.7446433
recall	macro	0.8223741	0.8575002	0.7572324
f_meas	macro	0.8558384	0.8885345	0.7474930

ISSN: 1477-9560