Attack detection in water distribution systems using machine learning

Table 2 A comparison of the algorithms on reduced BATADAL dataset

Rank	Name	No. attacks	S	\(S_{TTD}\)	\(S_{CLF}\)	\(F_1\)	TPR	TNR	PPV	TP	FP	TN	FN
1	QDA	7	0.9412	0.9448	0.9376	0.9032	0.8966	0.9786	0.9100	182	18	824	21
2	MD	7	0.8797	0.8536	0.9059	0.8877	0.8177	0.9941	0.9708	166	5	837	37
3	LOF	6	0.8553	0.8188	0.8918	0.7843	0.8696	0.9141	0.7143	180	72	766	27
4	Ensemble	6	0.8469	0.8245	0.8692	0.7211	0.8744	0.8640	0.6136	181	114	724	26
5	SOD	6	0.8284	0.8053	0.8515	0.6836	0.8621	0.8409	0.5663	175	134	708	28
6	OSVM	6	0.7824	0.7814	0.7833	0.6385	0.6700	0.8967	0.6099	136	87	755	67
7	Naive	7	0.7500	1.0000	0.5000	0.3261	1.0000	0.0000	0.1948	407	1682	0	0
8	LDA	4	0.6390	0.5588	0.7192	0.6096	0.4384	1.0000	1.0000	89	0	842	114