Table 3 Timeslice evaluation
From: Quantifying and filtering knowledge generated by literature based discovery
Filtering | Total | Correct | Precision | Recall | F-measure | Average |
|---|---|---|---|---|---|---|
Performance after a single step | ||||||
sy | 1,049,250,170 | 526,363 | 0.05 | 44.10 | 1.00e-03 | 11,131 |
manual | 386,952,997 | 268,327 | 0.07 | 22.48 | 1.38e-03 | 6,099 |
Y-Y&P | 243,218,893 | 190,072 | 0.08 | 15.93 | 1.56e-03 | 4,952 |
clt | 387,603,836 | 269,003 | 0.07 | 22.54 | 1.38e-03 | 6,103 |
break | 131,199,050 | 213,193 | 0.16 | 17.86 | 3.22e-03 | 2,232 |
Performance after two steps | ||||||
sy | 3,733,002,802 | 534,301 | 0.01 | 44.77 | 2.86e-04 | 39,602 |
manual | 1,638,685,466 | 274,544 | 0.02 | 23.00 | 3.35e-04 | 25,828 |
Y-Y&P | 994,744,004 | 194,749 | 0.02 | 16.32 | 3.91e-04 | 20,257 |
clt | 1,641,987,567 | 275,230 | 0.02 | 23.06 | 3.35e-04 | 25,857 |
break | 1,085,230,979 | 227,998 | 0.02 | 19.10 | 4.20e-04 | 18,467 |