Leveraging Gradient Boosting for Improved anti-HIV Activity Prediction
Main Article Content
Abstract
QSAR modelling is a widely used method that aims at learning relationships between input structures and output bioactivity data in order to make accurate predictions of bioactivities based on data structure. Prediction of the anti-HIV activity has been one of the most important tasks in chemical sciences where dominant approaches based on machine learning methods have been proposed. In this paper, we present a machine learning approach based on Gradient Boosting Regressor (GBR) to improve the performance of the HEPT anti-HIV activity prediction. The study was carried out with the estimation of the anti-HIV activity of a large set of 107 HEPT compounds using five quantum molecular descriptors. We evaluate our model on test and over all datasets, and in both cases we achieve state-of-the-art results.