Table 4 Predictive performance comparisons with different learning methods in Inter dataset validation.

From: Improved Classification of Blood-Brain-Barrier Drugs Using Deep Learning

Side Effects (SE)

Indications

SE + Indications

Training

Prediction

Training

Prediction

Training

Prediction

Deep Learning

0.975 ± 0.02

0.940 ± 0.04

0.987 ± 0.01

0.980 ± 0.02

0.971 ± 0.02

0.930 ± 0.06

0.940 ± 0.05

0.901 ± 0.07

0.982 ± 0.01

0.971 ± 0.25

0.965 ± 0.03

0.910 ± 0.08

0.930 ± 0.04

0.890 ± 0.09

0.941 ± 0.09

0.920 ± 0.06

0.931 ± 0.06

0.900 ± 0.09

  1. Each test data field shows average ± std of 1000 random splits of training and test data.