基于时间序列模型的传播预测方法研究  

Forecasting research based on time series analysis in communication

作　　者：陈秋怡 崔妮 汤景泰[3] CHEN Qiuyi;CUI Ni;TANG Jingtai(School of Journalism and Communication,Jinan University,Guangzhou 510632,China;School of Science and Engineering,Jinan University,Guangzhou 510632,China;School of Journalism,Fudan University,Shanghai 200433,China)

机构地区：[1]暨南大学新闻与传播学院,广州510632 [2]暨南大学理工学院,广州510632 [3]复旦大学新闻学院,上海200433

出　　处：《中国传媒大学学报（自然科学版）》2022年第3期57-65,共9页Journal of Communication University of China：Science and Technology

基　　金：国家社会科学基金重点项目“公共危机中的风险沟通与效果评估研究”(20AXW008);广东省“基于移动通信大数据的智慧交通分析、决策与优化平台”项目(2018A050501008)。

摘　　要：近年来,随着带有时间戳的在线数据的涌现以及计算方法的快速发展,为传播学突破传统横截式研究的局限性,在时间维度上进行扩展提供了契机。通过引入时间序列分析方法,可以实现纵向研究的三个目标:特征描述、变量预测与因果分析,本文研究了经典的时间序列预测方法,如ARIMA、SARIMA和动态回归模型在传播学的应用,并总结了实现预测分析的基础过程与原则。In recent years,the emergence of time-stamped online data and the rapid development of computational methods have provided an opportunity for communication studies to break through the limitations of traditional cross-sectional studies and expand in the time dimension.By introducing time series analysis methods,three objectives of longitudinal studies can be achieved:characterization,variable prediction and causal analysis.This paper investigates the application of classical time series forecasting methods,such as ARIMA,SARIMA and dynamic regression models in communication,and summarizes the process and principles underlying the realization of predictive analysis.

关 键 词：时间序列预测 ARIMA SARIMA 动态回归模型 传播学研究方法 

分 类 号：G203[文化科学—传播学]