摘要：A crucial part of drug development is the mechanism of action. It can assist scientists in the drug discovery process. This research presents a machine learning approach for predicting a drug's mechanism of action. Binary Relevance K Nearest Neighbors (Type A and Type B), Multilabel K-Nearest Neighbors, and a proprietary neural network are the machine learning models employed in this paper. The mean column-wise log loss is used to evaluate these machine learning models. With a log loss of 0.01706, the custom neural network model had the best accuracy. The Flask framework is used to integrate this neural network model into a web application. A user can upload a custom testing features dataset that includes gene expression and cell viability. The top drug classes will be displayed on the online application, along with scatter plots for each medication.
