Investigation of urinary system infection in pediatric patients with artificial intelligence methods

The detection of fungi, viruses, bacteria or other microorganisms in any part of the urinary system, which consists of the kidneys, bladder and urethra, is called a urinary system infection. Urinary system infections are more common in pediatric patients than in adults. In order to diagnose urinary system infections, doctors examine urine and blood tests, patient anamnesis and urine culture test results. The urine culture test, which is the basic diagnostic criterion, is concluded after 48-72 hours. The disease may worsen or antibiotic resistance may develop in patients by using unnecessary antibiotics until the urine culture test result is known. In this study, which investigates the possibility of early diagnosis of urinary system infections with artificial intelligence methods, two data sets were created by compiling data containing urine analysis, anamnesis, biochemistry, hemogram and urine culture test results of patients who applied to the Children's Polyclinic of Alaşehir State Hospital throughout 2023. The created data sets were tested with basic machine learning algorithms in the first part of the experiment and the results were compared according to classification metrics. In the second stage of the experiment, feature selection was performed using meta-heuristic optimization methods and the tests in the first stage were repeated for the resulting sub-data sets. The results obtained were examined comparatively and it was seen that machine learning algorithms provided superior performance after feature selection.