Comparative Analysis of ML Models for Symptom- Based Disease Prediction

This study puts emphasis on early diagnosis in healthcare under the motto "prevention is better than cure." It presents a machine framework for learning to anticipate diseases using patient-reported symptoms. The dataset used comprised 132 symptoms and 41 diseases from 4,921 patient records, and rigorous preprocessing and feature engineering ensured data quality. We implemented and optimized 6 machine learning models, including SVM, Logistic Regression, Random Forest, Gaussian Naive Bayes, K- Nearest Neighbors, and Decision Tree, achieving 100% classification accuracy and verified with precision, recall, and F1-score. The proposed system can easily be applied in real-world environments and even in those that have minimal resources due to scalability and computational effectiveness. The future work would extend the dataset, integrate the use of real-time monitoring, and validate among other demographics.