DrugPal: A Machine Learning Based Drug Recommender System To Assist Physician

The process of drug discovery has received a big boost from the use of high-end computers and machine learning algorithms. There are about 38 new drugs introduced into the market each year, making it difficult for physicians to keep up with the latest advancements. In this work we introduce an end-to-end system that is able to go through the literature of newly discovered drugs and store the features in a database. Then, it is able to accept the symptoms of a patient from a physician and recommend appropriate drugs, which may be prescribed to the patient, subject to the discretion of the physician. We use standard vectorization algorithms like word2vec and linearSVC to extract features from drug literatures. To prove the superiority of this method to other possible methods, we provide an ablation study, comparing the results of our method with the other most popular methods.