Sentiment Analysis of TikTok Netizens on Oil Palm Issues in Papua Using KNN

Abstract

Oil palm plantation in Papua has become a controversial issue that has generated diverse responses from the public, particularly regarding concerns over environmental impacts and ecosystem sustainability. These differing perspectives create the need to comprehensively and objectively understand public perceptions. This study aims to analyze the sentiment of Indonesian netizens toward the policy of oil palm plantation in Papua using a machine learning–based sentiment analysis approach. The research data were collected from user comments on the TikTok platform, which were subsequently processed through preprocessing stages, translation into English, and automatic labeling using TextBlob. The labeled data were then represented using Term Frequency–Inverse Document Frequency (TF-IDF) weighting and classified using the K-Nearest Neighbors (KNN) algorithm. The classification results using the K-Nearest Neighbors (KNN) algorithm indicate that out of a total of 220 data samples, 31 data (14%) were classified as positive sentiment and 189 data (86%) as negative sentiment. The classification process using the K-Nearest Neighbors (KNN) algorithm with the optimal K value of 5 achieved an accuracy of 77.27%, with a precision of 74.90%, recall of 66.00%, and an F1-score of 67.65%. The recall value for the positive class is relatively low, at 38.46%, indicating that the model still faces challenges in correctly identifying all positive data. This limitation is attributed to the imbalance in data distribution and the complexity of language used in social media comments. Nevertheless, the overall classification results suggest that the majority of netizens tend to oppose oil palm plantation in Papua, mainly due to concerns about environmental impacts and ecosystem degradation.