基于ARIMA时间序列模型的中医体质预测分析

Constitution Prediction of Traditional Chinese Medicine Based on ARIMA Time Series Model

目的:采用时间序列分析方法针对二十四节气中医体质类型变化构建ARIMA预测模型,探索中医体质预测研究新方法。方法:收集整理川西地区人群2020-2021年二十四节气内1 574例中医体质辨识数据作为有效样本,构建ARIMA时间序列模型,得到中医体质预测值与记录值的对应关系。结果:ARIMA (1,2,1)模型的预测拟合图显示预测数据与真实记录数据较为接近,可以较好地描述中医体质的变化趋势,模型预测精度较高,拟合优度良好。结论:ARIMA时间序列模型能够预测中医体质,为中医体质智能化提供新的研究思路。

Objective:The study aims to develop an ARIMA prediction model based on a time series analysis of constitutional type changes in Traditional Chinese Medicine(TCM) over 24 solar terms and to explore a new approach to TCM constitutional prediction.Methods:The ARIMA time series model is constructed by collecting and collating TCM constitutional identification data from 1574 patients in western Sichuan during the 24 solar terms of 2020 to 2021.The relationship between the predicted values and the recorded values of the TCM constitution is obtained to construct the model.Results:The ARIMA(1,2,1) model effectively predicted changes of TCM constitution.The prediction and fit plots show that the predicted data are close to the recorded data and the goodness of fit is high.Conclusion:The ARIMA time series model is capable of predicting changes in TCM constitutions so as to make dietary adjustment in advance and prevent the occurrence of diseases caused by the change of solar terms and provides a novel research direction for intelligent TCM constitution prediction.It is hoped that this approach will contribute to further understanding and development of TCM.