Fig. 11
From: Analysis of the benefits of imputation models over traditional QSAR models for toxicity prediction
Performance of models trained on training sets of increased sparsity. A line plot comparing performances of different techniques on the test set. Dotted lines indicate performances of models on original datasets (sparsity not increased). B scatter plot contrasting MCC scores of XGB-FN model for all individual assays. C scatter plot contrasting MCC scores of Macau model for all individual assays. Assays whose number of training points was unchanged (≤ 1000 data points originally) are colored red, those with reduced number of training points (> 1000 data points originally) are colored blue.