Sentiment analysis of Aceh tourism and its culture from Twitter data using support vector machine (SVM), naive Bayesian and k-nearest neighbour (KNN)
Reza Irwanda, Muslim Amiren, Zulfan Zulfan, Viska Mutiawani, Subianto, Rasuddin, Rini Deviani · AIP conference proceedings · 2024
Social media shares many opinions.Users can disseminate information, search for messages, announce activities, or communicate with others.A social media often used is Twitter.It includes opinion of the Indonesian people towards Aceh, especially regarding tourism and culture.To know that, the researchers conducted a sentiment analysis on Twitter using Support Vector Machine (SVM), Naive Bayessian and K-Nearest Neighbor (KNN).The steps taken are data collection, data preprocessing and analysis of classification results.The total data collected is 10,320 tweet data.Sentiment classification is divided into three categories, namely positive, negative and neutral.The accuracy of the model built with SVM is divided into three categories, namely positive-negative with an accuracy of 80.40%, neutral-negative with an accuracy of 81.11%, positive-negative with an accuracy of 76.50%.Using Naive Bayessian, model accuracy built is only 65.445% with a percentage split of 80 and K-Folds 10.The best model accuracy 71.3787% is built using K-Nearest Neighbor and K-value of 17, a percentage split of 80 and K-Folds 10.It results that related to Aceh's tourism and culture, negative sentiment is more than positive sentiment, 81.11% compared to 80.40%.Therefore, more efforts are needed to make Aceh's image more positive.