Drug Administration Route Classification using Machine Learning Models
G. Shobana, S. Nikkath Bushra · 2020
In the Pharmaceutical industries, the drug discovery process involves huge investment and is timeconsuming. The pre-clinical procedure may even take several years. There are several drug repositories that hold information about millions of drugs. Researchers around the globe continuously synthesize new compounds and explore its pharmaceutical activities. While its inhibition characteristics against diseases are diagnosed in the first phase, the second phase of testing would be the drug administration route. The route of administration like Oral, Parenteral, and Topical is identified at this stage. In the traditional manual investigation, only through biological lab testing, the administration route is determined. This paper investigates the performance of machine learning models in the classification of the drug administration route. Machine learning models like Logistic Regression, Random Forest, and Decision Tree were applied to the dataset and the effect of feature reduction on prediction accuracy was studied. Random Forest which is an ensemble learning model achieved high prediction accuracy.