基于支持向量机模型的河南艾滋病发病率预测

Forecast of incidence of AIDS in Henan Province based on support vector machine

摘要:

探索适合于河南省艾滋病发病趋势的预测模型，准确、快速地预测未来发病变化趋势，为制定艾滋病预防控制的策略和措施提供参考依据。方法收集河南省2000～2014 年艾滋病发病率数据，采用支持向量机模型建立其发病率预测模型。其中2000～2013 年发病率数据为训练样本，2014 年发病率数据为检验样本。以平均相对误差作为预测效果的评价指标。并用该模型对河南省2015～2019 年艾滋病的发病率进行预测。结果建立的支持向量机模型的平均相对误差为0.5512%。经预测，河南省2015～2019 年艾滋病的发病率分别为0.85/10 万、1.84/10 万、1.64/10 万、1.30/10 万、2.01/10 万。结论支持向量机模型有较高的预测精度及较小的预测误差，适用于河南省艾滋病的发病率预测。

Abstract:

To explore a model for forecasting acquired immunodeficiency syndrome (AIDS) in Henan Province, and accurately and quickly predicting the future trend of AIDS, so as to provide reference for AIDS prevention and control. Methods Data of AIDS incidence in Henan Province from 2000 to 2014 were collected. The incidence prediction model was established using support vector machine. The data from 2000 to 2013 were taken as training samples, and the data of 2014 were used as testing sample. Average relative error was used to evaluate the effect of prediction. Then the model was utilized to predict the incidence of AIDS in Henan Province from 2015 to 2019. Results The average relative error of the established support vector machine model was 0.5512%. It is predicted that the incidences of AIDS in Henan Province from 2015 to 2019 are 0.85/105, 1.84/105, 1.64/105, 1.30/105 and 2.01/105 respectively. Conclusions Support vector machine model has high prediction accuracy and small error, and is suitable for AIDS prediction in Henan Province.