AI-BASED DRUG DISCOVERY AND ACTIVITY PREDICTION

Drug discovery is a complex, expensive, and time-consuming process. This project presents
an AI-based drug discovery and activity prediction system that uses machine learning and
cheminformatics techniques to identify potential drug candidates efficiently. Molecular
compounds represented using SMILES notation are converted into molecular fingerprints and
analyzed using a Random Forest model to predict biological activity. The system is further
enhanced with multi-disease prediction, toxicity and side-effect analysis, and explainable AI
to improve prediction accuracy, safety assessment, and model transparency. An interactive
web interface allows users to perform real-time predictions and visualize results. This
intelligent framework reduces experimental effort, accelerates early-stage drug screening, and
supports safer and faster drug development.