Skip to main content

Table 2 Selecting DsIncr size for incremental model building (repeated 20 times, mean values reported)

From: Efficient iterative virtual screening with Apache Spark and conformal prediction

DsIncr (K)

Iterations

Accu.

Eff.

Docked mols (millions)

Total time (relative)

50

3.9

96.5

0.91

0.77

1

100

3.35

96.84

0.91

0.81

0.96

200

3.15

97.17

0.91

0.79

1.12

  1. Paremeters DsInit size = 200 K and Bins = 1_5 for all runs. Time was calculated relative to 50 K