Classification for Jaundice Symptoms Using Improved Dragonfly with XGBoost Model

The yellowing of the skin is a sign of neonatal hyperbilirubinemia, often known as jaundice. The most common method for screening for Hyperbilirubinemia in newborns is a visual examination; however, collecting blood specimens is the gold standard for identifying the condition and its harshness. Albumin-bound bilirubin was dissolved, producing the yellow hue. Too much haemoglobin breakdown results in bilirubin that the liver can't process. There are two ways for detecting bilirubin, an invasive one and a non-invasive one. Both methods have their advantages and disadvantages, but the invasive one needed more blood samples and was less pleasant to the patient. Neonatal jaundice is a physiological condition that affects about 80% of newborns. Because of this, hyperbilirubinemia occurs. Hence, it is important to check bilirubin levels on a frequent basis to determine whether patients require treatment and to identify those who may be at risk. The present continuous bilirubin measuring equipment is not only less precise than its predecessors but also more costly. In this research, we offer a diagnostic system for illness prediction using an enhanced version of the Extreme Gradient Boosting (XGBoost) classifier. A classifier's effectiveness relies heavily on the accuracy with which its hyperparameters are tuned. We employed the very effective hyper-parameter optimisation approach known as enhanced dragonfly optimisation (IDFO) on the XGBoost hyper-parameters. To liken the presentation of the suggested model to that of other replicas, tests are carried out.