Table 4.

Results from preliminary models on special care target (Top 10 of all models tested)

Model	Balanced accuracy	ROC AUC	F1 score	Time taken (s)
Bernoulli Naïve Bayes	0.90	0.90	0.92	0.14
QDA	0.88	0.88	0.91	0.22
Gausssian Naïve Bayes	0.85	0.85	0.95	0.15
xgBoost	0.85	0.85	0.96	1.31
LightGBM	0.82	0.82	0.96	0.47
AdaBoost	0.82	0.82	0.96	0.92
SVC	0.81	0.81	0.95	2.52
Random forest	0.81	0.81	0.96	1.14
Baging	0.80	0.80	0.96	0.78
Decision tree	0.80	0.80	0.96	0.23

Chosen algorithm for optimisation is highlighted