An Internet Public Opinion Monitoring Model Based on Stacked Ensemble Learning

With the rapid development of social media, internet public opinion has a profound impact on social stability and economic development. At present, common monitoring models generally have problems such as insufficient accuracy or weak generalization ability. Therefore, This paper proposes an internet public opinion monitoring model based on stacked ensemble learning, which aims to realize the real-time monitoring. The model selects combines random forest (RF), logistic regression (LR), support vector machine (SVM) group as the base learner. And integrates the prediction results of the base-learners through the meta learners (GBM) to monitor public opinion in real time. The experimental results show that the empirical model achieves 0.940 accuracy and an F1 score of 0.896, which significantly outperforms other ensemble methods. It can effectively monitor the scale and trend of public opinion of internet comments, and provide strong support for decision makers.