The healthcare data can be employed to develop a health prediction system that can improve in heart disease prevention. Big data on health care, including patient records, clinical notes, diagnosis, parents and family past ailments, hospitals, and scan results can aid in the phase of disease identification and prediction. The emerging machine learning method offers an important framework for forecasting cardiac diseases. An advanced Support Vector Machine (SVM) classifier was used by the program to conduct parameter tuning to improve classification accuracy and performance. The proposed work aims to develop a real-time prediction system for health issues based on big medical data processing on the cloud. In the proposed scalable system, the medical parameters are sent to Apache Spark to extract the attributes from the data and
to apply the proposed machine learning algorithm aiming to predict the healthcare risks and send them as alerts and recommendations to the users and the healthcare providers as well. The purpose of this paper is to evaluate the impact of applying machine learning algorithms using electronic health records. The proposed work aimed to provide an effective recommendation system using streaming medical data, historical data on the user profile, and knowledge database to provide users with the best recommendations and alerts in realtime according to the sensors measurements. The proposed system of prediction could offer high accuracy in comparison with literature work with the predictability of 90.6 for heart disease. The methodology of this research is applying parameterization for parameters on SVM to make the possibility of prediction is higher using the most effective features.