<p>At present, health prediction has become greatly significant in the medical domain. Healthcare prediction enhances diagnostic accuracy and aids in public health as well as medical management. Predictive analytics using big data allows researchers to develop different predictive models for forecasting health conditions and improving clinical outcomes. In recent years, cardiovascular disease has been considered to be the deadliest disease worldwide. Numerous research works have used both Machine Learning and Deep Learning approaches to attain superior decision-making processes. In this article, we propose a novel predictive framework named AdaBoost Convolutional-based Random Arithmetic Trigonometric algorithm for disease prediction. The proposed algorithm is designed for large datasets, offering probabilistic classification with feature independence. It is trained using cardiac disease data obtained from the Kaggle website. The dataset is utilized to accurately identify and classify both healthy and unhealthy cases. Also, it is suitable to validate with big data features and efficiency is estimated by validating with the determined dataset. Finally, the experimental evaluation is made in predicting the health condition of diverse patients. The results show that the proposed algorithm achieved a 98.85% accuracy rate, a precision of 98.21%, a recall of 98.76%, and an <i>F</i>1-score of 98.48%, enabling early disease detection and the prediction of patients’ future health.</p>