Analysis of risk factors for multi-drug resistant bacterial infections and prevention and control based on logistic regression analysis

Abstract Logistic regression and neural networks have developed rapidly in recent years, the poor diet of people modern society has led to emergence various diseases which drug-resistant bacterial infections occur during treatment, so this paper proposes logistic analyze risk factors preventive control multi-drug-resistant infections. A model was established determine magnitude effect each factor on dependent variable based standardized values, prevalence recoded prediction stage, with screened indicators serving as covariates. The number neurons input output layers is determined, weights are continuously adjusted iterations calculate average error rate between actual morbidities predicted values. gradient explosion dispersion problems in-depth analysis solved by selecting maximum probability for classification. values were calculated using cost function, adjusting parameters, comparing errors observed updating hidden layer thus improving accuracy analyzing multi-drug resistant preventing controlling deterioration disease. results showed that method, area ROC curve a discriminant, yielded an AUC 0.831 study, combined network predict bacteria higher 85.6%, identify potential occurrence, prevent aggravation infection.