Klasifikasi Kepuasan Guru SMA di Kabupaten Banyuwangi Terhadap Aplikasi Merdeka Mengajar Menggunakan Algoritma Naive Bayes

Classification of High School Teachers' Satisfaction in Banyuwangi Regency Towards the Merdeka Mengajar Application Using the Naïve Bayes Algorithm

Penelitian ini bertujuan mengidentifikasi indikator yang memengaruhi kepuasan guru SMA di Kabupaten Banyuwangi sebagai pengguna aplikasi Merdeka Mengajar, serta melakukan klasifikasi dan evaluasi menggunakan algoritma Naïve Bayes. Subjek penelitian adalah guru SMA di SMAN 1 Pesanggaran, SMAN 1 Bangorejo, dan SMAN 1 Glenmore. Data dikumpulkan melalui kuesioner berbasis model kepuasan pengguna DeLone dan McLean. Metode penelitian adalah survei kuantitatif. Analisis data menggunakan Information Gain dan Gain Ratio untuk seleksi fitur, diikuti penerapan algoritma Naïve Bayes. Hasil analisis menunjukkan bahwa indikator seperti Empathy (SEQ2), Availability (SQ2), dan Responsiveness (SEQ3) memiliki nilai Information Gain dan Gain Ratio tertinggi. Evaluasi model klasifikasi dengan confusion matrix dan metrik evaluasi (akurasi, presisi, dan recall) menunjukkan bahwa seleksi fitur Information Gain dan Gain Ratio meningkatkan kinerja model. Seleksi fitur Information Gain menunjukkan akurasi tertinggi sebesar 92.63%, presisi 98.18%, dan recall 90.31%, sementara seleksi fitur Gain Ratio menunjukkan akurasi 92.89%, presisi 98.75%, dan recall 90.31%. Simpulan dari penelitian ini adalah bahwa penggunaan seleksi fitur Information Gain dan Gain Ratio dapat meningkatkan kinerja klasifikasi Algoritma Naïve Bayes dalam mengukur kepuasan guru terhadap aplikasi Merdeka Mengajar. Indikator-indikator seperti Empathy, Availability, dan Responsiveness sangat penting dalam menentukan kepuasan pengguna. Saran untuk penelitian selanjutnya adalah untuk menguji model ini pada sampel yang lebih luas dan mengaplikasikan metode machine learning lainnya untuk membandingkan kinerja model.

Kata Kunci: Kepuasan Pengguna, Merdeka Mengajar, Klasifikasi, Naïve Bayes, Gain Ratio

This research aims to identify indicators that influence the satisfaction of high school teachers in Banyuwangi Regency as users of the Merdeka Mengajar application, and to classify and evaluate using the Naïve Bayes algorithm. The subjects of the study were high school teachers at SMAN 1 Pesanggaran, SMAN 1 Bangorejo, and SMAN 1 Glenmore. Data were collected through a questionnaire based on the DeLone and McLean user satisfaction model. The research method was a quantitative survey. Data analysis used Information Gain and Gain Ratio for feature selection, followed by the application of the Naïve Bayes algorithm. The analysis results showed that indicators such as Empathy (SEQ2), Availability (SQ2), and Responsiveness (SEQ3) had the highest Information Gain and Gain Ratio values. Model classification evaluation with confusion matrix and evaluation metrics (accuracy, precision, and recall) indicated that Information Gain and Gain Ratio feature selection improved model performance. Information Gain feature selection showed the highest accuracy of 92.63%, precision of 98.18%, and recall of 90.31%, while Gain Ratio feature selection showed an accuracy of 92.89%, precision of 98.75%, and recall of 90.31%. The conclusion of this study is that the use of Information Gain and Gain Ratio feature selection can improve the classification performance of the Naïve Bayes algorithm in measuring teacher satisfaction with the Merdeka Mengajar application. Indicators such as Empathy, Availability, and Responsiveness are crucial in determining user satisfaction. Recommendations for future research include testing this model on a larger sample and applying other machine learning methods to compare model performance.

Keywords: User Satisfaction, Merdeka Mengajar, Classification, Naïve Bayes, Gain Ratio