Journal of Big Data

Table 6 ROC AUC values computed on different benchmark datasets using State-Of-Art algorithms and the proposed schemes

From: Unsupervised outlier detection in multidimensional data

Dataset	State-of-art												Proposed
Dataset	KNN	KNNW	LOF	SLOF	LoOP	LDOF	ODIN	FABOD	KDEOS	LDF	INFLO	COF	BADk	Scheme 3
Arrhythmia	0.7930	0.7674	0.7674	0.7500	0.7500	0.7551	0.7663	0.7715	0.6680	0.8565	0.7950	0.7663	0.8074	0.8770
Heartdisease	0.8644	0.8311	0.8533	0.7800	0.7977	0.7577	0.8544	0.8911	0.6844	0.8933	0.8333	0.8955	0.8467	0.9400
Hepatits	0.8955	0.8706	0.9452	0.8806	0.8905	0.8706	0.9228	0.8408	0.8706	0.9403	0.9005	0.8955	0.8393	0.9403
Parkinson	0.9895	0.9895	0.9895	0.9895	0.9895	0.9895	0.9895	0.9895	0.9895	1	0.9895	0.9791	0.9583	1
Spambase40	0.5734	0.5661	0.4738	0.5011	0.4965	0.4796	0.5191	0.4372	0.4766	0.5364	0.4738	0.4994	0.6034	0.6125
Glass	0.8748	0.8832	0.8666	0.8650	0.8395	0.7788	0.7292	0.8579	0.7420	0.9035	0.8037	0.8953	0.9122	0.9293
Pendigit	0.9868	0.9854	0.9168	0.9107	0.9039	0.7181	0.9225	0.9610	0.8113	0.9545	0.8908	0.9609	0.9826	0.9819
Shuttle	0.9890	0.9861	0.9896	0.9869	0.9869	0.9775	0.9888	0.8381	0.9810	0.9922	0.9863	0.9920	0.9865	0.9895
WBC	0.9929	0.9920	0.9906	0.9835	0.9737	0.9582	0.9563	0.9892	0.6023	0.9929	0.9821	0.9863	0.9842	0.9989
WPBC	0.5409	0.5319	0.5254	0.5018	0.5018	0.5034	0.5072	0.5341	0.5185	0.5829	0.4957	0.5568	0.5427	0.5801

Bold values indicate are best-achieved

Back to article page