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

Table 3

SigNCD versus the baselines on Chinese Finance News.

Algorithms	Prec.	Rec.		Runtime (ms)
SigNCD w/ P1	0.98	0.97	0.98	970
NCD	0.97	0.88	0.92	154487
SpotSigNCD w/ P1	0.98	0.80	0.88	7187
SpotSigs	0.97	0.90	0.93	8823
Google’s simhash	0.99	0.94	0.96	13151
SL+ST	0.94	0.61	0.74	8353481