A Sentiment Classification Model of E-Commerce User Comments Based on Improved Particle Swarm Optimization Algorithm and Support Vector Machines

With the rapid increase of number Internet users and amount online comment data, a large referable information samples are provided for data mining technology. As technical application mining, text sentiment classification can be widely used in public opinion management, marketing, other fields. In this study, combination approach to SVM (support vector machine) IPSO (improved particle swarm optimization) is proposed classify by using data. First, 30,000 goods reviews corresponding ratings collected through web crawler. Then, TFIDF (term frequency-inverse document frequency) Word2vec vectorize review Next, model trained SVM, initial parameters optimized IPSO. Finally, we applied SVM-IPSO test set evaluated performance several measures. Our experiment results indicate that performed best classification. Additionally, traditional machine learning becomes very effective after parameter optimization, which demonstrates parameters’ optimization has successfully improved accuracy. Furthermore, our significantly outperforms benchmark models, indicating it could improve accuracy efficiency