Figure 5 | Scientific Reports

Figure 5

From: Predicting the outcomes of organic reactions via machine learning: are current descriptors sufficient?

Figure 5

Classification based on chemical-linguistic descriptors. (a) An example showing two organic molecules and their maximal common substructure – such substructures computed over millions of molecule-molecule pairs can be used as chemical-linguistic descriptors, CLDs. (b) Examples of some smaller and larger CLDs used as descriptors to predict reaction yields and times. Dashed lines denote aromatic bonds. (c) Performance of a random forest classifier based on various numbers of CLDs. Even for 5,000 descriptors, the misclassification error is still ca. 40%.

Back to article page