论文部分内容阅读
目的建立一种基于支持向量机的麻疹发病率预测模型,为麻疹的预防控制决策提供参考依据。方法收集我国麻疹1996-2015年发病率数据,以1996-2014年发病率数据为训练样本,以2015年发病率数据为检验样本,采用支持向量机回归算法建立我国麻疹年发病率的预测模型,并对我国2016-2018年的麻疹发病率进行预测。结果经该模型的预测,麻疹发病率的预测值和实际值吻合度较高,平均相对误差为0.620 07%。我国2016-2018年麻疹的发病率分别为3.23/10万、3.13/10万、3.79/10万。结论采用支持向量机回归模型预测我国麻疹的年发病率是可行有效的。
Objective To establish a prediction model of measles morbidity based on support vector machine, and provide reference for prevention and control decision of measles. Methods The data of incidence of measles in China from 1996 to 2015 were collected. The incidence data of measles from 1996 to 2014 were taken as training samples. The incidence of measles in 2015 was used as the test sample. The predictive model of annual incidence of measles in China was established by using support vector machine regression algorithm. And forecast the incidence of measles in China from 2016 to 2018. Results The prediction of the model showed that the predicted value of measles incidence was in good agreement with the actual value, with an average relative error of 0.620%. The incidence rates of measles in China from 2016 to 2018 were 3.23 / 100,000, 3.13 / 100,000 and 3.79 / 100,000, respectively. Conclusion Prediction of annual incidence of measles in China by using support vector machine regression model is feasible and effective.