Analisis Sentimen terhadap Berita Palestina: Pendekatan Regresi Logistik terhadap Dataset Multi-Platform Sosial Media

Abstrak

This study aims to analyze user sentiment toward news related to the Palestinian issue and to determine its impact on the popularity of news stories across various social media platforms. Additionally, this study aims to develop a news popularity prediction model using logistic regression by incorporating text representation features based on Term Frequency–Inverse Document Frequency (TF-IDF). The data used were sourced from the UCI Machine Learning Repository, which contains metrics on user interactions and news content across several social media platforms. The research stages include selecting Palestine-related news articles, text preprocessing, sentiment analysis using a lexicon-based approach, TF-IDF feature extraction from news headlines and summaries, the creation of popularity labels based on engagement scores, and modeling using logistic regression. The results of the study show that the sentiment variable has a significant influence on the probability of a news story going viral. In addition, the inclusion of TF-IDF features improves the model’s accuracy and Area Under the Curve (AUC) compared to the baseline model without text features. Based on these findings, it can be concluded that user sentiment and text representation play a crucial role in predicting the virality of news regarding the Palestinian issue on social media. These findings contribute to our understanding of information dissemination patterns surrounding humanitarian issues and can serve as a foundation for future information management systems and public opinion analysis.