基于量子遗传算法优化的新Prophet模型及其验证  

A New Prophet Model Based on Quantum Genetic Algorithm and Its Verification

作　　者：邓翔[1] 彭杰 吕一清[1] DENG Xiang;PENG Jie;LV Yi-qing(School of Economics,Sichuan University,Chengdu 610065,China)

机构地区：[1]四川大学经济学院,四川成都610065

出　　处：《系统工程》2020年第5期141-150,共10页Systems Engineering

基　　金：教育部人文社科基金资助项目(17YJC790104);成都科技厅软科学项目(2019-RK00-00048-ZF);四川大学中央高校基本科研业务费项目(2019自研-经济017)。

摘　　要：时间序列预测是预测领域中的一个重要研究方向,尤其是对受节假日、季节性以及突发事件冲击较大的时间序列进行预测仍然存在着一些缺陷,表现在拟合效果差、精确度低和不稳定性方面。本文在Faccbook设计的Prophet模型基础上,构建一种以X-13-ARIMA-SEATS模型得到超参数的经验范围,利用量子遗传算法优化超参数的新Prophet模型,该模型能够更好地拟合和预测含有季节性和突发事件冲击的时间序列,并且大大简化超参数的优化。进一步,采用宏观时间序列CPI作为实证对象,通过新Prophet模型对其进行预测并与其他时间序列模型的预测效果对比,其实证结果对新Prophet模型的预测效果进行了佐证。Time series prediction is an important research direction in the field of forecasting,especially for the time series impacted by holidays,seasonality and emergencies,there arc still some defects,such as poor fitting,low prediction accuracy and instability.Based on the prophet model designed by Facebook,this paper constructs a new Prophet model,which uses x-13-arima-sets model to get the empirical range of hyper parameters,and optimizes the hyper parameters by using quantum genetic algorithm.The model is more effective in fitting and predicting time series with seasonal and unexpected events,and greatly simplifies the optimization of hyper parameters.Furthermore,we take the macro time series CPI as the empirical object and predict it using new prophet model.Compared with other time series models,the prediction results of the new prophet model is better which verifies the advantagesof the new model.

关 键 词：CPI 超参数优化 X-13-ARIMA-SEATS 新Prophet 

分 类 号：F224[经济管理—国民经济]