Analisis sentimen masyarakat terhadap PON 2024 Aceh dan Sumatra Utara di Twitter menggunakan algoritma SVM dan KNN

Abstrak

The 2024 National Sports Week (PON) held in Aceh and North Sumatra is one of the largest sports events that attracted public attention on social media, particularly on Twitter. This study compares the performance of Support Vector Machine (SVM) and K-Nearest Neighbor (KNN) in sentiment analysis related to PON 2024. The collected data consists of 1,458 tweets, including 1,321 positive tweets and 137 negative tweets, which were processed using text pre-processing techniques. The results show that SVM with TF-IDF achieved the highest accuracy (95.45% on the training data and 91.78% on the testing data), with an F1-score of 96%, demonstrating the most stable performance. Meanwhile, KNN performed best at K = 4 with Word Embedding, achieving 95.20% accuracy on the training data and 92.12% on the testing data, but exhibited greater performance variation than SVM. This study highlights that both SVM and KNN are reliable for sentiment analysis on social media, making a significant contribution to the evaluation of public perception of PON 2024.