Comparative Analysis of Machine Learning Algorithms for Multiple Disease Prediction Model with An Optimized Scalable Deployment

Non-communicable diseases (NCDs) like diabetes, heart disease, Parkinson's disease, and breast cancer are leading causes of global mortality and morbidity. Early detection is crucial to improve outcomes and reduce healthcare costs, but traditional diagnostic methods are often inaccessible in low- resource settings. This study leverages machine learning (ML) to develop a unified Multiple Disease Prediction System using Logistic Regression, Random Forest, and Support Vector Machines (SVM). Datasets were pre-processed to handle missing data and ensure feature scaling, resulting in high accuracy, recall, and F1-scores across diseases. The system achieved its best performance with Random Forest and SVM algorithms. Random Forest produced consistent accuracy for Parkinson's disease (94.87%) and breast cancer (98.83%), whereas SVM performed well with high-dimensional data sets like breast cancer (98.25%) and Parkinson's disease (94.87%). Logistic Regression, though simpler, demonstrated competitive performance for diabetes (75.32%) and heart disease (85.25%). The models were deployed via a simple-to-use Streamlit web app, enabling real-time, accessible medical forecasts providers and individuals. The research identifies concerns such as dataset heterogeneity and computational complexity, and proposing future directions such as real-time data integration, cloud-based scalability, and explainable AI. This article demonstrates ML's ability to revolutionize early disease diagnosis and enhance global health access.