Analisis Sentimen Terhadap RUU TNI Di Platform X (Twitter) Menggunakan Metode Ensemble Berbasis Naive Bayes Dan Support Vector Machine 

Sentiment Analysis On The Indonesian National Army (Tni) Bill On Platform X (Twitter) Using An Ensemble Method Based On Naive Bayes And Support Vector Machine 

Revisi Rancangan Undang-Undang Tentara Nasional Indonesia (RUU TNI) menjadi isu strategis yang memicu dinamika opini publik di media sosial. Penelitian ini bertujuan untuk menganalisis sentimen masyarakat terhadap Revisi RUU TNI di platform X (Twitter) serta menelaah bagaimana isu tersebut dibingkai oleh media daring nasional. Sebanyak 15.090 tweet dianalisis menggunakan model klasifikasi Naive Bayes, Support Vector Machine (SVM), dan pendekatan Ensemble Weighted Soft Voting (Naïve Bayes–SVM). Hasil klasifikasi menunjukkan bahwa opini publik didominasi oleh sentimen negatif sebesar 47,87%, diikuti sentimen netral sebesar 33,87%, dan positif sebesar 18,26%. Model SVM menghasilkan akurasi sebesar 95,34%, diikuti Naïve Bayes dengan 93,67%, sementara ensemble Weighted Soft Voting mencapai 96,00%, meskipun peningkatan tersebut tidak signifikan secara statistik. Analisis framing terhadap tiga media besar menunjukkan perbedaan dalam membingkai isu: Detik.com menonjolkan sudut pandang negatif terhadap pemerintah, Kumparan.com bersikap netral–negatif dengan menyoroti keresahan warga, dan Kompas.com cenderung netral dengan nada kritis terhadap potensi kemunduran demokrasi. Hasil penelitian i ini menegaskan bahwa sentimen publik di ruang digital terhadap Revisi RUU TNI cenderung negatif dan menunjukkan kekhawatiran terhadap meningkatnya peran militer dalam ranah sipil, sekaligus menggarisbawahi pentingnya komunikasi publik yang transparan dan pemberitaan media yang berimbang.

The Revision of the Indonesian National Armed Forces Bill (RUU TNI) has become a strategic issue that triggers dynamic public opinion on social media. This study aims to analyze public sentiment toward the revision of the RUU TNI on platform X (Twitter) and to examine how the issue is framed by national online media. A total of 15,090 tweets were analyzed using three classification models: Naive Bayes, Support Vector Machine (SVM), and an Ensemble Weighted Soft Voting approach (Naïve Bayes–SVM). The classification results show that public opinion is dominated by negative sentiment at 47.87%, followed by neutral sentiment at 33.87%, and positive sentiment at 18.26%. The SVM model achieved accuracy of 95.34%, followed by Naïve Bayes with 93.67%, while the ensemble Weighted Soft Voting model reached 96.00%, although the improvement was not statistically significant. The framing analysis of three major media outlets revealed differences in how the issue was portrayed: Detik.com emphasized a negative perspective toward the government, Kumparan.com adopted a neutral–negative stance highlighting public concerns, and Kompas.com maintained a neutral tone with a critical view of potential democratic regression. Overall, the findings indicate that digital public sentiment iii toward the RUU TNI revision tends to be negative, reflecting concerns over the growing role of the military in civilian affairs. The study underscores the importance of transparent public communication and balanced media reporting on sensitive national defense issues.