论文部分内容阅读
目的探讨气象因子对脑出血发病的影响及其相关性,并建立基于气象因子的脑出血发病的预报模型,为预防和控制脑出血发病提供参考依据。方法收集2008-2010年荆门市脑出血逐日病例数和同期气象资料,应用SPSS 13.0软件,分析气象因子与脑出血发病的相关性,构建多元线性回归和前馈型神经网络(BP神经网络)模型预测脑出血发病人数,并对模型进行评价。结果气象因子与脑出血发病密切相关。线性回归模型为y赞=18.248-0.873x(1x1为气温,r=0.708,F=18.13,P<0.01),BP神经网络模型结构为7-6-1,2种模型预测的平均误差率分别为16.15%和8.08%,非线性相关系数分别为0.795和0.907。结论荆门市脑出血发病与气温呈显著负相关,BP神经网络模型有较好的预测效果。
Objective To explore the impact of meteorological factors on the incidence of cerebral hemorrhage and its correlation, and to establish a prediction model of cerebral hemorrhage based on meteorological factors, and provide a reference for the prevention and control of cerebral hemorrhage. Methods The number of daily cases of intracerebral hemorrhage and meteorological data in Jingmen City during 2008-2010 were collected. The correlation between meteorological factors and the incidence of cerebral hemorrhage was analyzed by using SPSS 13.0 software to build a model of multiple linear regression and feedforward neural network (BP neural network) Predict the number of patients with cerebral hemorrhage, and evaluate the model. Results Meteorological factors and cerebral hemorrhage are closely related. The linear regression model was y = 18.248-0.873x (1x1 is the temperature, r = 0.708, F = 18.13, P <0.01). The BP neural network model structure was 7-6-1. The average error rates of the two models were 16.15% and 8.08% respectively. The nonlinear correlation coefficients were 0.795 and 0.907 respectively. Conclusions There is a significant negative correlation between the incidence of cerebral hemorrhage and temperature in Jingmen City. The BP neural network model has good predictive value.