Enhancing the prediction of vitamin D deficiency levels using an integrated approach of deep learning and evolutionary computing

Vitamin D deficiency (VDD) has emerged as a serious global health concern that can lead to far-reaching consequences, including skeletal issues and long-term illness. Classical diagnostic approaches, although effective, often include invasive techniques and lacks to leverage the massive amount of healthcare data. There is an increasing demand for noninvasive prediction approaches for determining the severity of VDD. This work proposes a novel approach to detect VDD levels by combining deep learning techniques with evolutionary computing (EC). Specifically, we employ a hybrid deep learning model that includes convolutional neural networks (CNN) and bidirectional long short-term memory (BiLSTM) networks to predict VDD data effectively. To improve the models effectiveness and guarantee the optimal choice of the features and hyper-parameters, we incorporate evolutionary computing methods, particularly genetic algorithms (GA). The proposed method has been proven effective through a comprehensive assessment on a benchmark dataset, with 97% accuracy, 96% precision, 97% recall, and 96% F1-score. Our approach yielded improved performance, when compared to earlier methods. This research not only push forward predictive healthcare models but also shows the potential of merging deep learning with evolutionary computing to address intricate health-care issues.