Sentiment Classification of Social Media User Comments Using SVM Models

Text sentiment analysis is often used for online public opinion analysis and prediction. This study evaluates the performance of the Support Vector Machine (SVM) by benchmarking it against four traditional classifiers: LR, KNN, NB, and XGBoost, and analyses the precision rate, recall rate, and F1 score of each model. The results indicate that SVM model training is more effective in obtaining the sentiment classification model in this paper than the other four models. This study proposes the use of machine learning algorithms to classify the sentiment of comments related to ‘Huawei mate60’ on Little Red Book. This will enable Huawei and other brands to monitor and analyse consumer sentiment on social media platforms in an automated way. The proposed approach can help brands respond to consumer feedback in a timely manner and provide data support for product improvement, marketing strategy development, and brand image management.