Minería de datos aplicada a la información médica relacionada con infecciones urinarias

Abstract

Abstract: Up to five years of age, urinary tract infections in children are one of the main reasons for hospitalization and consultation, it is necessary to diagnose, control and provide timely medications, making use of medical information from a local hospital, analyzes variables for one hundred and seventy-two patients, where fields such as age, sex, germ, bacterial count (cfu), sensitivity, resistance to twenty-nine medications among others. Some algorithms such as K-means, K-modes, hierarchical and DBSCAN based on three main methods prior to the application of the algorithms like dimensionality reduction using Binary PCA, similarity matrices and the use of complete information. Algorithms evaluated by the Silhouette coefficient allowed us to find interesting information regarding girls up to 12 months old. age, because although they had performed a urine test and there was a presence of bacteria in their urine cultures did not reflect a specific germ, on the other hand, sensitivity to Ampicillin and Sulbactan as well as resistance to Ertapenem are present in the majority of groups analyzed.