The Design Of Medical-Meteorological Forecasting Model For Cardiovascular Disease Based On Artificial Neutral Network

OBJECTIVE:At present, with the acceleration of urbanization, the increasing of urban population density, the quickening pace of life, the changing dietary structure and environmental factors, some cardiovascular disease, such as coronary heart disease, myocardial infarction and arrhythmia, seriously threat to people's physical health, especially to old people. A lot of studies suggest that the pathogenesis of cardiovascular disease has close relation with some certain meteorological conditions, so the disease forecasting according to meteorological data attracts more and more attention. This study aims to identify the key meteorological indices which cause cute outbreak of cardiovascular disease, analyze the pathogenesis, establish the medical meteorological forecasting model based on artificial neural network (ANN), so that we can provide scientific reference for medical-meteorological forecasting of cardiovascular disease.METHODS:1. The data of cardiovascular disease from January 2007 to December 2008 were collected in 20 emergency stations of Emergency Centre of Jinan, at the same time the data of meteorological factors were obtained from China Meteorological Data Sharing Service System. The relationship between meteorological indices and the number of cardiovascular disease in every 10 days was analyzed by using SAS9.1 in order to find the meteorological indices that have bigger relevance cardiovascular disease and statistical significance.2. Establish the artificial neural network structure, the transfer function and training methods of forecasting model, build and evaluate the model by using MATLAB7.0.3. Establish the interface of forecasting system by using LabWindows/CVI, build the server and client of forecasting system and realize network operation and individualized warning.RESULTS:The outbreak of cardiovascular disease has close relationship with temperature, pressure, humidity, and the changing extent of various meteorological factors. The results show that cardiovascular disease easily break out when the weather is becoming dry and cold, and the temperature is sharply descending. The result of forecast shows that forecast accuracy rate can reach 91%. The medical-meteorological forecasting system can make out different levels of medical-meteorological warning and individualized treatment according to the result of forecast. The system can be operated through a point-to-point network.CONCLUSION:The medical-meteorological forecasting model for cardiovascular disease based on BP neutral network shows good prediction effect and further research value. The forecasting system realizes network operation, individualized warning and has higher prediction accuracy, strong practicability and simple operation. When applied in medical-meteorological forecast, it can effectively avoid the ourbreak of cardiovascular disease caused by the bad weather, improve the quality of life, reduce health-care spending and the burden on society.