Multiple Disease Prediction System Using Machine Learning
Rahul Shukla, Rupali Sawant
Abstract
Machine learning (ML) refers to the science and engineering of artificially intelligent systems, providing them with the capability to learn without being explicitly programmed. In recent years, ML in the healthcare domain has made great advancements in the early predictions of many critical illnesses. While there have been significant contributions to single disease prediction systems (like one for heart disease prediction, one for diabetes, and so on), there lacks the existence of an integrated regulatory framework that performs the prediction of multiple diseases under a single interface. This paper proposes a novel “Multiple-disease Prediction System” for the early-stage prediction of Heart Diseases, Parkinson's Disease, Breast Cancer, and Diabetes using efficient Machine Learning algorithms. The novelty of this work lies in the implementation of multiple ML classifiers for each prediction system, followed by their performance evaluation and comparison based on their respective test accuracies. From experimental results, it was observed that Logistic Regression (LR) and Support Vector Machine (SVM) outperformed Decision Tree and Random Forest classifiers in all cases. The highest reported accuracy for LR was 77% for the diabetes model, while SVM obtained accuracy values of 80%, 92%, and 97% for heart disease, Parkinson's disease, and breast cancer models, respectively. Additionally, the work employs efficient data preprocessing techniques with feature selection carried out via Principal Component Analysis (PCA), and by implementing correlation matrices among independent variables. This resulted in reducing the number of features while still maintaining the test accuracies of individual models for the correct prediction of multiple diseases—a step not efficiently captured by previous works. Finally, the integrated framework is deployed in the form of a web app (with multiple constraint checks) to enhance the overall user experience.