| Summary: | The explosion of text-based communication among internet users in Malaysia has created a dynamic online landscape for political discourse. As most internet users engage in social activities and discussions, it becomes imperative to analyze public sentiments towards political issues. This study presents a comparative analysis of sentiment analysis techniques using both traditional ML (SVM, NB, RF, LR) and advanced DL methods for political issues in Malaysia. This study aims to explore the capability of the hybrid DL algorithm and ML algorithm in the sentiment classification of political views based on Twitter data. The data has undergone preprocessing, including data cleaning, normalization, and feature extraction. Then, the data was labelled to ensure accuracy and relevance. Finally, modelling techniques were applied to analyze and derive insights from the prepared dataset. The hybrid model PSO-CNN has performed better in the balanced dataset with oversampling implementation by the Random Oversampling technique. The results have been divided into the data exploratory and the algorithm’s performance analyses. Based on the performance analysis, hybrid DL PSO-CNN-based sentiment analysis has proven to be more efficient and can classify positive and negative tweets with an acceptable accuracy of 81%. The future work includes experimenting with other hybrid CNN algorithms and testing the PSO algorithm with other ML. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.
|
|---|