The interest of coumarins as antioxidant agents has attracted much attention in recent years. A quantitative structure-activity relationship (QSAR) study of the DPPH• (2,2-diphenyl-l-picrylhydrazyl) radical scavenging ability of chemical compounds, based on the 0-3D DRAGON molecular descriptors and an artificial neural networks (ANN) technique was developed. The built mathematical model showed a correlation coefficient for the training set (R2) = 0.71, an external correlation coefficient ( and it was used to predict the antioxidant activity of 4-hydroxycoumarin derivatives. Besides, an experimental in vitro assay was developed for the reference compound of this group (4-hydroxycoumarin) and the results obtained confirmed the predictions made by the ANN.
This is a very interesting work, using ANN for modeling of DPPH• Free Radical Scavenging Activity of Coumarin Derivatives. In this sense I have a question. Do you consider the use of other machine learning techniques? Why do you choose ANN in first place?
for model validation. Other machine learning techniques, such as Boosting, Support Vector Machines, Gaussian process, etc. are to be considered in future works.
Stephen jones barigye
