<![CDATA[In this study, the LinearSVC algorithm from the Support Vector Machine (SVM) method was used for sentiment analysis on IMDb review data. Feature extraction was carried out using the TF-IDF Vectorizer. The main challenge lay in determining the value of the hyperparameter, particularly the regularization parameter (C), which greatly influences the quality of the prediction results. To address this issue, the study employed the Particle Swarm Optimization (PSO) algorithm to find the optimal value of C. Experiments showed that without optimization, the SVM model achieved an accuracy of only 89.48%, but after applying PSO, the optimal C value of 0.1612 was found, which increased the model’s accuracy to 92.03% on the test data. Additionally, other evaluation metrics also showed significant improvements, with a Precision of 91.29%, Recall of 92.92%, and F1-Score of 92.10%. The significance of this improvement indicates that the PSO method consistently outperforms conventional approaches that rely on manual hyperparameter selection or grid search, which are often slower and less accurate in finding the optimal value. The results of this study demonstrate that hyperparameter optimization using PSO can significantly enhance SVM performance in sentiment classification. This approach is not only relevant for analyzing IMDb reviews but can also be applied to various other NLP tasks, such as public opinion analysis and product reviews, making it an efficient solution for improving text classification accuracy.]]>
