关键词: 循环系统疾病, 疾病谱, 气象诱因, 机器学习, 预报模型

Abstract:

Objective： Study the incidence characteristics and the meteorological triggers of circulatory system diseases in Panzhihua， construct a risk level prediction model， in order to provide disease prevention services for the government，hospitals and other departments， as well as the general public in Panzhihua.Methods： The medical case data certain stage of circulatory system diseases and daily meteorological data during the same period in Panzhihua Central Hospital were collected and organized. Based on the basis of exploring the spectrum and temporal variation characteristics as well as the meteorological triggers of local circulatory system diseases， the prediction model for the risk level of circulatory system diseases in Panzhihua was constructed by the machine learning algorithms， and the prediction effect was tested.Results and Conclusion： The most affected population was the elderly， followed by middle-aged people for the circulatory system diseases in Panzhihua. The number of male patients was about 1.7 times that of females. And the highest number of cases occurred in winter， followed by summer， and the lowest in autumn. January was the peak and october was the valley， that was different from the eastern part in China， the number of outpatients was most significantly affected by the lowest temperatures， with a corresponding increase in the number of outpatients when the lowest temperatures drop to lower. The accuracy of BP forecast model trial forecast was 89.96%. The forecast accuracy of ELMAN forecast model was 93.61%. ELMAN forecasting model was better than BP in forecasting the change trend and series of risk grade，and the model was more stable.

Key words:  , Circulatory system disease； , Disease spectrum； , Meteorological triggers； , Machine learning； , Forecasting model