Performance Comparison of KNN and NB Algorithms in Predicting Urinary Tract Infection in Pediatric Patients

Infections developing in any part of the urinary system, which consists of the kidneys, urethra and bladder, are called urinary system infections. Although urinary system infections are also seen in adults, they are more common in pediatric patients. In the clinical evaluation of urinary system infe...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:2025 International Conference on Emerging Systems and Intelligent Computing (ESIC) S. 653 - 657
Hauptverfasser: Gundogdu, Huseyin, Gundogdu, Tuba, Altay, Osman
Format: Tagungsbericht
Sprache:Englisch
Veröffentlicht: IEEE 08.02.2025
Schlagworte:
Accuracy
Bayes methods
Classification algorithms
Classification Evaluation Metrics
Diseases
Distance measurement
K-Nearest Neighbor Algorithm
Machine learning
Machine learning algorithms
Machine Learning Classification Algorithms
Naive Bayes
Nearest neighbor methods
Prediction algorithms
Urban areas
Urinary Tract Infections
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Infections developing in any part of the urinary system, which consists of the kidneys, urethra and bladder, are called urinary system infections. Although urinary system infections are also seen in adults, they are more common in pediatric patients. In the clinical evaluation of urinary system infections, doctors examine urinalysis, patient anamnesis and urine culture test results. The most important indicator, urine culture test, results are obtained after 48-72 hours, which makes it difficult to decide on treatments. In this study, a data set was created by compiling data containing urine analysis, anamnesis and urine culture test results of patients who applied to the Pediatric Outpatient Clinic of Alaşehir State Hospital throughout 2023. The obtained data set includes 20 different features of 759 different patients. To predict the disease before the urine culture test results, K-Nearest Neighbor and Naive Bayes were used from machine learning classification algorithms. The used algorithms were evaluated with different parameters and their performances were analyzed according to classification metrics. The K-Nearest Neighbor algorithm achieved the highest accuracy rate of 90.73 when the city block distance measurement method and the k value were taken as 5.
DOI:10.1109/ESIC64052.2025.10962604