Sentiment analysis of wayang climen using naive bayes method

Abstract

This research focuses on sentiment analysis of Wayang Climen performances in Indonesia using the Naïve Bayes algorithm. Wayang, a traditional puppet show, holds cultural significance and has persisted alongside modern entertainment options. The study collected public comments from Dalang Seno and Ki Seno Nugroho's YouTube channels, classified them into positive, negative, and neutral sentiments, and employed a translation process to align comments with program language objectives. Preprocessing steps included case folding, removing punctuation, tokenizing, stopword removal, and post-tagging. To address data class imbalances, resampling was performed using the Synthetic Minority Oversampling Technique (SMOTE). The Naïve Bayes algorithm was utilized for data classification, exploring various translation scenarios. Evaluation involved the confusion matrix method and metrics like accuracy, precision, recall, and f-measure. Results demonstrated that the Dalang Seno train data scenario outperformed Ki Seno Nugroho's, with higher precision, recall, accuracy, and f-measure values. Additionally, the translation scenario from Indonesian to English yielded the most effective results. In conclusion, this study highlights the suitability of the Naïve Bayes algorithm for sentiment analysis in the context of Wayang Climen performances, with practical implications for understanding public sentiment in the digital age.