Optimized disease prediction in healthcare systems using HDBN and CAEN framework

Classification and segmentation play a pivotal role in transforming decision-making processes in healthcare, IoT, and edge computing. However, existing methodologies often struggle with accuracy, precision, and specificity when applied to large, heterogeneous datasets, particularly in minimizing false positives and negatives. To address these challenges, we propose a robust hybrid framework comprising three key phases: feature extraction using a Hybrid Deep Belief Network (HDBN), dynamic prediction aggregation via a Custom Adaptive Ensemble Network (CAEN), and an optimization mechanism ensuring adaptability and robustness. Extensive evaluations on four diverse datasets demonstrate the framework’s superior performance, achieving 93 % accuracy, 87 % precision, 95 % specificity, and 91 % recall. Advanced metrics, including a Matthews Correlation Coefficient of 0.8932, validate its reliability. The proposed framework establishes a new benchmark for scalable, high-performance classification and segmentation, offering robust solutions for real-world applications and paving the way for future integration with explainable AI and real-time systems.

• •
  Designed a novel hybrid framework integrating HDBN and CAEN for adaptive feature extraction and prediction.
• •
  Proposed dynamic prediction aggregation and optimization strategies enhancing robustness across diverse data scenarios.