Data Science-Driven Toxicology Prediction: Severity Classification and Treatment Recommendation Using Machine Learning

Toxicology-related cases are a growing concern in India, with frequent incidents of pesticide poisoning, drug overdoses, and envenomations. The major challenges include delayed diagnosis, lack of structured toxicology databases, limited accessibility to poison control centers, and the absence of real-time decision-making tools. This leads to high fatality rates, especially in rural areas where specialized toxicology units are scarce. To address these challenges, this project applies data science and machine learning to automate poisoning case identification, symptom-based severity classification, and treatment recommendations. The system uses Natural Language Processing (NLP) for fuzzy search, TF-IDF vectorization for feature extraction, and a Random Forest Classifier for severity prediction. By leveraging structured data from open-source medical repositories, hospital records, and poison information centers, this model enhances toxicology decision-making and helps medical professionals provide faster, data-driven responses to poisoning emergencies.