A Unified Framework for Behaviour Monitoring and Abnormality Detection for Smart Home

Table 6

Detailed accuracy statistics.

	Model	Accuracy	FP Rate	Precision	Recall	F-Measure	Kappa statistic	PRC	ROC
Morning Results (5am-12pm)	SVM	92.33%	0.003	0.984	0.983	0.981	0.9082	0.977	0.990
	MLP	97.67 %	0.003	0.979	0.977	0.977	0.9723	0.986	0.998
	KNN	92.81 %	0.011	0.930	0.928	0.927	0.9144	0.967	0.995
	J48	94.96%	0.008	0.950	0.950	0.949	0.9399	0.959	0.992
	RepTree	91.72%	0.014	0.918	0.917	0.915	0.9010	0.996	0.970
	Naïve Bayes	79.91%	0.029	0.811	0.799	0.791	0.7610	0.981	0.888
Afternoon Results (12pm-5pm)	SVM	95.32%	0.007	0.951	0.954	0.952	0.9438	0.945	0.987
	MLP	99.92%	0.001	0.999	0.999	0.999	0.9991	1.000	1.000
	KNN	88.59%	0.006	0.965	0.965	0.965	0.8620	0.989	0.998
	J48	93.99%	0.011	0.941	0.940	0.946	0.9278	0.994	0.968
	RepTree	97.42%	0.02	0.974	0.974	0.973	0.9690	0.999	0.987
	Naïve Bayes	84.30%	0.015	0.880	0.843	0.850	0.8152	0.934	0.986
Evening Results; (5pm-8pm)	SVM	95.35 %	0.003	0.954	0.953	0.949	0.9440	0.949	0.992
	MLP	98.44%	0.000	0.998	0.999	0.999	0.9810	1.000	1.000
	KNN	96.51 %	0.006	0.966	0.965	0.965	0.9580	0.988	0.998
	J48	80.62%	0.040	0.809	0.806	0.805	0.7660	0.806	0.994
	RepTree	82.62%	0.260	0.886	0.826	0.829	0.7920	0.973	0.862
	Naïve Bayes	93.30%	0.007	0.912	0.993	0.921	0.9190	0.999	0.978
Night Results; (8pm-5am)	SVM	98.86 %	0.001	0.990	0.989	0.989	0.9864	0.998	1.000
	MLP	99.97%	0.000	1.000	0.999	1.000	0.9997	1.000	1.000
	KNN	96.55 %	0.006	0.966	0.966	0.960	0.9580	0.987	0.998
	J48	96.54%	0.002	0.959	0.969	0.905	0.9510	0.986	0.956
	RepTree	90.39%	0.013	0.912	0.904	0.905	0.8850	0.986	0.980
	Naïve Bayes	94.89%	0.006	0.935	0.949	0.941	0.9390	0.999	0.988