Navigating Uncertainty: Advancing Disease Prediction with Semi-Supervised learning

The study introduces a novel semi-supervised machine learning approach for disease diagnosis, utilizing complete blood count (CBC) report data. Recognizing the complexity of disease overlap, the framework embraces uncertainty in predictions, mirroring real-world diagnostic challenges. Comparative analysis with conventional supervised machine learning algorithms underscores the efficacy of the proposed technique. Through meticulous preprocessing and feature extraction from CBC data, supervised models effectively classify diseases such as anemia, nutritional deficiencies, and infections. Leveraging unlabeled data enriches predictions with nuanced insights, albeit with inherent uncertainty. Experimental validation demonstrates the approach’s effectiveness in automating disease diagnosis, while adeptly navigating diagnostic intricacies, thus offering a promising avenue for improving healthcare outcomes.