MUMBAI, India, Aug. 31 -- Intellectual Property India has published a patent application (202641101616 A) filed by Vignan'S Nirula Institute Of Technology Science For Women on August 22, 2026, for Heart Disease Risk Prediction Using Hybrid Machine Learning Approaches.

Inventors include B Aruna; K. V. S. S. Rama Krishna; V. Pavani; P. Sandhya Krishna; K Poojitha; L Pranavi; and P Divya.

The application for the patent was published on August 28, 2026, under issue no. 35/2026.

Abstract: This invention is about a Hybrid Machine Learning System and Method for Early Risk Prediction of Heart Disease. With this system, the diagnosis of cardiovascular diseases can be achieved with greater accuracy, reliability, and efficiency. This invention collects the clinical information of patients from medical databases and electronic health records, and uses rigorous data preprocessing techniques to improve data quality. The system generates engineered clinical parameters such as body mass index (BMI), mean arterial pressure (MAP), and pulse pressure (PP), and uses statistical feature selection methods like ANOVA, Chi-square, and mutual information. In order to deal with the imbalance of classes, the Synthetic Minority Oversampling technique (SMOTE) was used to help create synthetic samples of the minority class to improve performance. The processed dataset was put through a hybrid stacked ensemble framework as Random Forest, Support Vector Machine, Logistic Regression, XGBoost, and Multi-Layer Perceptron and the outcomes have been merged using a meta-learner. The invention includes Platt Scaling to allow for adjustment through calibration of probabilities, decision making supported with confidence and Explainable Artificial Intelligence techniques to help create transparent predictions and ensure trustworthiness of results. Evaluation of the framework demonstrates the fact that it has yielded better results compared to conventional machine learning methods in line with accuracy, precision, recall, F1 score, ROC AUC, robustnes

Disclaimer: Curated by HT Syndication.