Predicting the Toxicity of Chemicals and Drugs using Machine Learning Models
Abstract: Toxicity is the way to find out if the drug/medicine is harmful to human body. Currently, the toxicity of the medicine is calculated using in-vivo method, where the medicine is tested on the animals and their results are generated. However, this method of toxicity testing for all existing compounds biologically may not be viable financially and logistically. We try to solve this problem by using machine learning and deep learning techniques. We have used the ensemble learning algorithm voting based classifier [logistic regression, decision tree, support vector machines] to predict the toxicity of theTox21 dataset. Where we get the AUC (Area under Curve) of NR-AR-LBD: 0.87 SR-mmp: 0.84 NR-Ahr: 0.81 on these assays.
Keywords: Machine Learning, Ensemble Learning, Toxicity Prediction, Chemicals, Tox21 dataset
How to Cite?: Shraddha Surana, Pratik Mahankal, Prateek Bihani, "Predicting the Toxicity of Chemicals and Drugs using Machine Learning Models", Volume 10 Issue 4, April 2021, International Journal of Science and Research (IJSR), Pages: 1110-1114, https://www.ijsr.net/getabstract.php?paperid=SR21423181604, DOI: https://dx.doi.org/10.21275/SR21423181604