Analisis Sentimen Siswa Terhadap Fasilitas Pembelajaran Jurusan TKJ Menggunakan Algoritma Naïve Bayes

Sari

Abstrak: Pendidikan kejuruan merupakan sistem pendidikan yang berorientasi pada kesiapan kerja, salah satunya melalui Sekolah Menengah Kejuruan (SMK). Jurusan Teknik Komputer dan Jaringan (TKJ) di SMKN 1 Kota Bima menuntut kualitas pembelajaran yang didukung fasilitas dan strategi pembelajaran yang efektif. Penelitian ini bertujuan menganalisis sentimen siswa terhadap pembelajaran TKJ menggunakan algoritma Naïve Bayes. Kebaruan penelitian terletak pada penerapan analisis sentimen berbasis Naïve Bayes pada data opini siswa TKJ di SMKN 1 Kota Bima yang sebelumnya belum pernah dilakukan, serta penggunaan data primer langsung dari siswa untuk mengevaluasi kualitas pembelajaran dan fasilitas sekolah. Dibandingkan metode lain seperti Support Vector Machine (SVM) dan Deep Learning, Naïve Bayes dipilih karena lebih sederhana, efisien, dan mampu bekerja baik pada jumlah data terbatas. Data penelitian diperoleh melalui penyebaran kuesioner kepada siswa kelas X–XII dengan total 106 data. Tahapan penelitian meliputi preprocessing, pembobotan TF-IDF, dan pengujian menggunakan confusion matrix. Hasil penelitian menunjukkan akurasi sebesar 77,27%, dengan performa terbaik pada kelas positif, sedangkan kelas netral memiliki performa terendah akibat jumlah data yang sedikit.
Kata kunci: Naïve Bayes, Analisis Sentimen, Data Mining

Abstract: Vocational education is an education system focused on work readiness, one of which is through Vocational High Schools (SMK). The Computer and Network Engineering (TKJ) program at SMKN 1 Kota Bima requires high-quality instruction supported by effective facilities and teaching strategies. This study aims to analyze student sentiment toward TKJ learning using the Naïve Bayes algorithm. The novelty of this research lies in the application of Naïve Bayes-based sentiment analysis to data on the opinions of TKJ students at SMKN 1 Kota Bima, which has not been done before, as well as the use of primary data directly from students to evaluate the quality of learning and school facilities. Compared to other methods such as Support Vector Machine (SVM) and Deep Learning, Naïve Bayes was chosen because it is simpler, more efficient, and capable of performing well with limited data. Research data was obtained through the distribution of questionnaires to students in grades X–XII, yielding a total of 106 data points. The research stages included preprocessing, TF-IDF weighting, and testing using a confusion matrix. The results showed an accuracy of 77.27%, with the best performance in the positive class, while the neutral class had the lowest performance due to the small amount of data.
Keywords: Naïve Bayes, Sentimen Analysis, Text Mining