基于百度指数的突发事件网络舆情预测分析

Prediction and Analysis of Network Public Opinion in Emergencies Based on Baidu Index

摘要: 如何在公共卫生事件发生时，正确引导和管理网络舆情，形成积极健康的舆情氛围，一直是相关部门需要关注的问题。本文以上海新冠疫情为对象，爬取网络舆情相关的百度搜索指数样本数据(2022.11.10~2023.01.31)。首先，依据舆情生命周期理论，将该阶段上海新冠疫情网络舆情发展划分为萌芽期–成长期–成熟期–衰退期，对不同时期网络舆情的传播特点进行了定性分析，并给出各阶段舆情管理建议。然后，考虑到样本数据的非线性和非平稳的特点，分别利用最小二乘支持向量机(LSSVM)和自回归移动平均模型(ARIMA)进行预测。最后，将LSSVM和ARIMA两个模型结果赋以适当权重进行组合预测，相较于之前单一模型预测，预测精度明显提升。本文对突发公共事件网络舆情进行预测分析，了解舆情发展趋势，以期为相关部门引导和管控网络舆情提供一定的参考依据。

Abstract: How to correctly guide and manage online public opinion in public health emergencies to foster a positive and healthy atmosphere has always been a concern for relevant departments. This article focuses on the COVID-19 epidemic in Shanghai and crawls sample data from the Baidu search index related to online public opinion, spanning from November 10, 2022 to January 31, 2023. Firstly, based on the life cycle theory of public opinion, the article classifies the development of online public opinion into four stages: Germination, growth, maturity, and decline. A qualitative analysis of the propagation characteristics of online public opinion in these different periods is conducted, followed by suggestions for public opinion management in each stage. Then, considering the nonlinear and non-stationary characteristics of the sample data, the least squares support vector machine (LSSVM) and autoregressive moving average model (ARIMA) are utilized for prediction. Finally, the results of the LSSVM and ARIMA models are combined with appropriate weights to enhance prediction accuracy, significantly outperforming previous single model predictions. By predicting and analyzing the development trend of public opinion during sudden public events, this article aims to provide a reference for relevant departments in guiding and controlling online public opinion.