Intelligent Machine Learning-Based System for Disease Prediction and Drug Recommendation

Abstract

The development of automated systems that may assist in clinical diagnosis and therapy selection has been prompted by the increasing accessibility of digital medical records and patient-generated data. An intelligent machine learning-based system for precise illness prediction and tailored medication prescription is presented in this work. To improve decision reliability, the suggested methodology combines sentiment-enriched medication review analysis with structured symptom datasets. Based on user-provided symptoms, many supervised learning models are trained to detect likely illnesses, such as Multinomial Naïve Bayes, Decision Tree, Extra Tree Classifier, and Support Vector Machine. In parallel, patient drug evaluations are subjected to natural language processing methods including VADER sentiment scoring and TF-IDF vectorisation, which provide weighted sentiment metrics that inform medication recommendations. The ensemble-enhanced prediction system exhibits good accuracy and resilience across many clinical settings, according to experimental assessment. The integrated strategy guarantees that clinical characteristics and patient experience both contribute to more individualised, well-informed, and comprehensible healthcare suggestions.