Hybrid RF-DT Model for Chronic Disease Detection Using EHR Big Data Management and Analytics

Abstract

Big data EHR aggregates voluminous data of patients and, after processing, presents valuable insights that aid in the detection and management of chronic diseases. This paper, therefore, describes a hybrid model for chronic disease detection using combined Random Forest and Decision Tree models. Chronic diseases such as diabetes, cardiovascular diseases, and hypertension remain the major health burdens worldwide, and their early detection would be an important aspect in their management. Traditional methods of detection are based on very limited and sporadic data collection, hence the scope for real-time and timely decisions remains highly inhibited. The model proposed here effectively integrates large-scale EHR data and, by finding unobvious patterns and dependencies in the patient records, results in more accurate and reliable forecasting. The experimental results proved that the RF+DT model outperformed state-of-the-art techniques concerning accuracy, precision, recall, and F1-score and turned out to be much robust in chronic disease detection. This approach offers further computational efficiency, hence feasibility for real-time healthcare applications. The results of this work will go toward the enhancement of machine learning applications in healthcare, offering a scalable and efficient framework for the detection of chronic diseases that can be further optimized for a variety of healthcare environments.