Measles Outbreak To Public Health Emergencies Grading Quantified To Determine The Modeling And Evaluation

Background and objectives Infectious diseases has seriously threatened human health, timely detection and management of infectious diseases public health emergencies can effectively reduce the spread and harm. Currently, in the Chinese public health emergencies network direct reporting system, the sudden public health incidents of infectious diseases still rely on manual classification completed, which can not meet public health emergencies rapid response requirements, an urgent need to develop a set of information and computer resources of automatic decision-making system to support scientific decision-making. To achieve that goal, the judge need to select appropriate indicators and quantitative grading, related to the establishment of the discriminant model.Content and Methods The issue of public health emergency reporting system in 2005-2007 measles-related emergencies, the use of Clementine 11.1.1 software, the method of combining C5.0 decision tree model and more orderly classification of the response variables logistic regression analysis of measles-level events to determine that the laws that determine choice of indicators and quantitative grading, and establish levels for the computer automatically determine the best model.Results Through C5.0 algorithm , the decision tree model and variable indicators regrouping standards are produced. By more orderly classification of the response variables logistic regression, measles-related emergencies were predictable logistic regression model. Training data back and predicting the test data are used to evaluate the decision tree model of the C5 and the logistic regression model .The result shows that the C5 decision tree model training to the correct rate of 92. 57%, test data correctly forecast rate of 85. 71%; the logistic regression model training data back to the correct rate of 94. 06 %, the forecast accuracy rate of the test data of 83. 67%; the accuracy of the two models to determine and forecast the level of measles-related emergencies is high, and the two-level model with the original judge that the level of the inspection results with a high degree of consistency, the two models predict no significant difference.Conclusion Through the research, the specific forecasting indicators and division standards of measles-related emergencies are identified, combined with decision tree model and logistic regression model can be realized on the measles-automatic emergency purposes. In this study, the method and process automatically provide reference for other infectious diseases and the other public health event discriminant method.