Efficient Development of Supervised Learning Algorithm for kidney Stone Prediction

Many individuals visit the emergency room due to severe discomfort caused by kidney stones. A kidney stone is a hard, solid particle of a substance that forms as a result of the minerals in the urine. They are the outcome of a combination of genetic and environmental factors. Obesity, certain foods, drugs, and a lack of water can all contribute to this illness. Kidney stones can be of various shapes and sizes. Numerous research and imaging modalities have detected the existence of kidney stones. To completely interpret and diagnose, these pictures require the knowledge of a medical specialist. Clinicians who use computer-aided diagnosis systems as supplemental tools gain considerably from the practical techniques provided by these systems. Nonetheless, the presence of noise has resulted in some errors in kidney stone classification. Machine learning (ML) has recently made remarkable progress, and this study suggests an automatic diagnosis of kidney stones based on coronal computed tomography scans (CT). For research, cross-sectional CT images were collected, yielding a total of 2600 images. Random Forest (RF), Logistic Regression (LR), and an ensemble (combination of the LR and RF) are used. The different metrics are used to compare the ML model for identifying the best model. From the metrics comparison, it is identified as all the positive metrics will be high for the ensemble method and negative metrics will be low for the LR model.