PERBANDINGAN ALGORITMA SVM DAN NAIVE BAYES PADA ANALISIS SENTIMEN TERHADAP KEBIJAKAN PENGHAPUSAN KEWAJIBAN SKRIPSI DI SOSIAL MEDIA X

Rani Yunita; Mia kamayani;

ABSTRAK

Perbandingan Algoritma SVM dan Naive Bayes Pada Analisis Sentimen Terhadap Kebijakan Penghapusan Kewajiban Skripsi Di Sosial Media X

Pada Agustus 2023 Kementerian Pendidikan, Kebudayaan, Riset, dan Teknologi membuat peraturan
salah satunya menghapus kewajiban skripsi sebagai syarat kelulusan di semua perguruan tinggi di
Indonesia. Pro dan kontra saling bertukar tempat di berbagai media sejak peraturan menteri tersebut
diumumkan. Banyak yang mendukung kebijakan tersebut tetapi tidak sedikit yang menentang. Dari
Issue tersebut, peneliti melakukan analisis sentimen di X tentang kebijakan yang menghapus
kewajiban skripsi sebagai syarat kelulusan menggunakan 700 data. Penelitian ini membandingkan
hasil evaluasi algoritma Support Vector Machine (SVM) dengan Naïve Bayes menggunakan 80%
data latih dan 20% data uji. Berdasarkan hasil yang diperoleh dari penelitian ini didapatkan 331
sentimen positif serta 369 sentimen negatif dan ditarik kesimpulan bahwa Support Vector Machine
(SVM) menjadi algoritma yang terbaik dengan accuracy 80%, recall 83%, precision 76%, dan F1-
Score 79%.

Kata kunci: Analisis Sentimen, X, Skripsi, Support Vector Machine (SVM), Naïve Bayes

Comparison of SVM and Naive Bayes Algorithms on Sentiment Analysis of Thesis Obligation Removal Policy on Social Media X

In August 2023, the Ministry of Education, Culture, Research and Technology made a regulation,
one of which eliminated the obligation to write a thesis as a graduation requirement in all
universities in Indonesia. Pros and cons have been exchanging places in various media since the
ministerial regulation was announced. Many support the policy but not a few are against it. From
this issue, researchers conducted sentiment analysis on X about the policy that removed the thesis
obligation as a graduation requirement using 700 data. This research compares the evaluation
results of the Support Vector Machine (SVM) algorithm with Naïve Bayes using 80% training data
and 20% test data. Based on the results obtained from this study, 331 positive sentiments and 369
negative sentiments were obtained and it was concluded that Support Vector Machine (SVM) was
the best algorithm with 80% accuracy, 83% recall, 76% precision, and 79% F1-Score.

Keywords: Sentiment Analysis, X, Thesis, Support Vector Machine (SVM), Naïve Bayes