Analysis of Multi-Class Sentiment on Indonesian Twitter Using Support Vector Machine Classification Algorithm with Particle Swarm Optimization.

With the rapid evolution of the internet, widespread interactions occur across diverse media platforms, including Twitter. Tweets shared on Twitter often need to be categorized into different sentiment polarities, often referred to as multi-class classification. This research aims to enhance the accuracy of sentiment classification utilizing the Support Vector Machine (SVM) algorithm through the application of Particle Swarm Optimization (PSO). PSO is employed to determine optimal hyperparameter values that maximize the performance of the SVM classification model. Two models were employed for testing: SVM classification alone and SVM classification integrated with PSO optimization. The accuracy of the SVM classification yielded 63.7%, while the combined SVM-PSO approach achieved 65.8%. Consequently, this approach led to a 2.1% accuracy enhancement following optimization through the PSO algorithm.