Effective and Fast Near Duplicate Detection via Signature-Based Compression Metrics

Table 2

SigNCD versus the baselines for Gold Set.

Algorithms	Prec.	Rec.		Runtime (ms)
SigNCD w/ P1	0.95	0.89	0.92	7838
SpotSigNCD w/ P1	0.97	0.87	0.92	10853
NCD	0.90	0.78	0.83	56829
SpotSigs	0.90	0.77	0.83	12824
Google’s simhash	0.85	0.45	0.59	13010
SL+ST	0.74	0.53	0.62	242383