JoWUA

Abstract

In the field of biomedicine, machine learning (ML)-based algorithms are an effective tool widely used to benefit from the control of diseases in their early stages. As the body expels urea daily, measuring its level is crucial for controlling proper renal function and identifying potential kidney disorders at an early stage. Currently, urea-detecting methods have involved a long duration, high costs, and a large workforce. For rapid, accurate, and economical detection, an optimized algorithm has been developed for urea detection using a machine learning approach. The motivation of the current work is to take advantage of machine learning-based algorithms in the context of urea reports in biomedical applications. The proposed algorithm would utilize several machine learning methodologies, including Support Vector Machine, Artificial Neural Network, and Random Forest, to process data collected from an extensive array of different biomedical sensors. This will enable the algorithm to be deployed on more diverse and complex datasets, allowing for more accurate prediction of urea levels. The algorithm is further developed through training with a cream pie of urea measurements that incorporate more patient-specific data. Finally, the obtained results will be compared to those from the well-established traditional pyridol methods to detect urea, testing the efficiency of the developed method. Thus, it is expected that the new algorithm for the optimized devices will perform better than conventional methods by yielding more accurate and up-to-date results for urea detection.