Skip to main content

Table 2 Comparison of methods across five different datasets using fivefold cross validation

From: DRABAL: novel method to mine large high-throughput screening assays using Bayesian active learning

Method GMean (%) F1Score (%) F0.5Score (%)
RandomOrder-10 51.14 36.84 46.44
DRABAL 61.05a 51.11a 64.52a
  1. The HTS assays data is partitioned into five approximately equally sized mutually distinct subgroups such that a single subgroup representing 20% of the data is retained for testing only. For each partition (fold) of the data, the model is developed on the training portion and evaluated on the testing portion. The results from the testing folds are averaged to produce an estimation of performance. Statistically significant difference when compared with all other methods over fivefolds using t-test at the 5% significance level is denoted by a