Table 5.

Validation of machine learning classification with new data set (n = 45, 22 healthy controls, 8 slight, 15 moderate or severe chronic periodontitis patients).

Group		Model	Feature combination	Accuracy	Balanced accuracy	Sensitivity	Specificity
1	H vs. M-S	Neural Network	Tf + Pg + Pi + Fn + Pa + Cr + Td	0.84	0.85	0.93	0.77
	H vs. M-S	Random Forest	Tf + Pg + Fn + Td+ Ec + Cr	0.86	0.88	0.93	0.82
	H vs. M-S	Support Vector Machine	Tf + Pg + Pi + Pa + Td	0.78	0.81	0.93	0.68
	H vs. M-S	Regularized Logistic Regression	Tf + Pg + Pi + Cr	0.81	0.83	0.93	0.73
			Average	0.82	0.84	0.93	0.75
2	H vs. Sli-M-S	Neural Network	Tf + Ec + Pg + Pa + Td	0.69	0.69	0.74	0.64
	H vs. Sli-M-S	Random Forest	Tf + Ec + Aa + Pg + Pa + Cr	0.76	0.75	0.78	0.73
	H vs. Sli-M-S	Support Vector Machine	Tf + Cr + Pa + Pi + Fn	0.67	0.66	0.87	0.45
	H vs. Sli-M-S	Regularized Logistic Regression	Tf + Cr + Pg + Pa + Aa + Fn + Pi	0.71	0.71	0.83	0.59
			Average	0.71	0.70	0.80	0.60
3	H vs. Sli	Neural Network	Tf	0.63	0.51	0.25	0.77
	H vs. Sli	Random Forest	Tf + Pg	0.73	0.66	0.50	0.82
	H vs. Sli	Support Vector Machine	Tf +Td	0.63	0.51	0.25	0.77
	H vs. Sli	Regularized Logistic Regression	Tf + Aa + Td	0.60	0.53	0.38	0.68
			Average	0.65	0.55	0.34	0.76