| Summary: | 碩士 === 國立雲林科技大學 === 工業工程與管理系 === 107 === Urinary Tract Infection is one of the most common bacterial infections in infants and young children. Asymptomatic fever is the frequent symptoms of children or toddlers with urinary tract infection. Diagnosing Urinary Tract Infection is difficult particularly in young infants or children because they could not explain their feeling clearly. If medical personal could not detect and treat properly, childhood UTI lead to renal scarring, hypertension, and end-stage renal disease. Febrile children when in suspicion of UTI need to have urine testing which is collected by either catheterization of the urethra or suprapubic aspiration, both of which can be viewed as relatively invasive and time-consuming procedures. Because of the current diagnostic and predictive techniques for urinary tract infections, the complexity and variability of the cause of the disease is hard to under control effectively.
Therefore, this research through all available clinical information to build the predict model, that combine genetic algorithms with Naïve Bayes Classifiers and Neural Network, for the diagnosis of a UTI. And the impact factors of affecting UTI were gender, born weight, C-reactive protein, white blood cell and bacterial. We design an application software to access of the risk of urinary tract infections in infants and young children with fever. In case of a baby with fever that cannot express clearly, the nursing staff can use the application software to screen the symptoms, and we can know the chance of having a urinary tract infection from this results.
According to possibilities of UTI, it is necessary to evaluate whether to do the next urine screening. Not only the evaluation time of diagnosis but also the unnecessary waste of medical resources can be reduced.
|
|---|
