摘要
利用主成分分析(PCA)对影响因子进行筛选后,再利用LSSVM进行建模和预测,由此得到了一种新的时间序列预测模型(PCA-LSSVM)。以甘肃省天水市的胆囊炎发病率为例,应用PCA-LSSVM预测模型进行实证分析,结果表明,此模型的预测精度优于PCA-MLR,ARIMA,LSSVM等参比模型。
After the principle components analysis（PCA）is used to filter the impact factors and the prediction model is conducted by LSSVM, a new time series prediction model is obtained（PCA- LSSVM）. Taking the cholecystitis incidence rate of Tianshui in Gansu province as an example, the empirical analysis is conducted using PCA-LSSVM prediction model. The results show that PCA-LSSVM＇s performance is superior to the reference models of PCA-MLR, ARIMA, LSSVM.
出处
《唐山学院学报》
2013年第3期8-11,22,共5页
Journal of Tangshan University
基金
国家自然科学基金资助项目(60673192)
四川省科技厅资助项目(2013JY0125)
关键词
最小二乘支持向量机
主成分分析
时间序列
预测
胆囊炎发病率
least square support vector machine~ principal component analysis~ time series~ fore- casting cholecystitis incidence rate
