基于PCA-ARIMA模型的高血压发病率预测

目的高血压发病率是政府和相关医学工作者预防和监测高血压的重要依据之一。方法利用主成分分析(principal component analysis,PCA)对因子进行线性筛选,获得保留因子后利用ARIMA进行建模预测,即为PCA-ARIMA多维时间序列组合预测模型。结果高血压发病率的拟合与独立预测结果表明,PCA-ARIMA优于PCA-MLR、ARIMA等参比模型。结论本文提出的基于主成分分析和ARIMA模型(PCA-ARIMA模型)的建模有助于提高模型的预测精度。

Prediction of Hypertension Incidence Rate Based on PCA-ARIMA Model

Objective Hypertension incidence rate is an important basis for government and related medical workers to prevent and to survey hypertension.Methods This paper proposed a combination prediction model(PCA-ARIMA),i.e.prediction model of PCA-ARIMA multidimensional time series combination,in which the impact factors are filtered by principal component analysis(PCA),and then the model were established based on ARIMA.Results The results of fitting of hypertension incidence rate and independent prediction showed that PCA-ARIMA’s performance was superior to reference models.Conclusion The proposed model based on the principal component analysis and ARIMA model could help improve prediction precision of the model.