Table 5

Modeling phases and data partitions

Phase

Partitions/activity

Oral Clearance

LD50

Cytotoxicity


Phase I

training set

348 of 435 (80%)

3,095 of 3,869 (80%)

4,000 of 8,983 (44%), chosen from 70%

testing set

87 of 435 (20%)

774 of 3,869 (20%)

2,695 of 8,983 (30%)

Phase II

training set

435 of 435 (100%)

3,869 of 3,869 (100%)

4,500 of 8,983 (50%), includes all positive labels


Boik and Newman BMC Pharmacology 2008 8:12   doi:10.1186/1471-2210-8-12

Open Data