Skip to main content


Fig. 7 | Journal of Cheminformatics

Fig. 7

From: KekuleScope: prediction of cancer cell line sensitivity and compound potency using convolutional neural networks trained on compound images

Fig. 7

RF and ConvNet predictions on the test set. Correlation between the predictions for the test set compounds calculated with RF models and ConvNets trained on the same training set instances. Overall, the predictions show a positive and significant correlation (P < 0.05; Pearson’s correlation coefficient values in the 0.72–0.84 range). The predictions for the ten runs are shown

Back to article page