Analisis Kesesuaian Komentar Mahasiswa Pada Sistem Akademi Online Angket Penilaian Dosen Menggunakan Supervised Model

Sigit Wijanarko (STMIK Antar Bangsa)

Jurnal Teknik Informatika STMIK Antar Bangsa
Vol 8 No 2 (2022): JTI Periode Agustus 2022

Abstract—The existence of an online academic system of lecturer assessment questionnaire in the Faculty of Computer Science - Budi Luhur University can help the management and lecturers in seeing the lecturer performance index based on student assessment. The questionnaire includes the level of student satisfaction in the form of a Likert scale numerical score against 19 (nineteen) categories, and student comments in natural language which can contain criticisms and suggestions for lecturers. Problems arise in classifying student comments because the writing style is free and not rigid, the online academic system of lecturer assessment questionnaire has not been able to predict the sentiment of student comments and correlation with the category of lecturer assessment. This study tries to classify comments in Indonesian language for comment analysis, which is correlated with the category’s lecturer assessment. It is hoped that the results of this study can help the university management and lecturers know the correlation between numerical scores and student comments. Keyword—Text classification, Natural Language, Naïve Bayes, K-Nearest Neighbor, Support Vector Machine. Intisari—Adanya sistem angket penilaian dosen sistem akademik online di Fakultas Ilmu Komputer - Universitas Budi Luhur dapat membantu pihak manajemen dan dosen dalam melihat indeks kinerja dosen berdasarkan penilaian mahasiswa. Kuesioner tersebut meliputi tingkat kepuasan mahasiswa berupa skor numerik skala Likert terhadap 19 (sembilan belas) kategori, dan komentar mahasiswa dalam bahasa natural yang dapat berisi kritik dan saran bagi dosen. Permasalahan muncul dalam pengklasifikasian komentar mahasiswa karena gaya penulisannya yang bebas dan tidak kaku, sistem akademik online angket penilaian dosen belum mampu memprediksi sentimen komentar mahasiswa dan korelasinya dengan kategori penilaian dosen. Penelitian ini mencoba mengklasifikasikan komentar dalam bahasa Indonesia untuk analisis komentar, yang dikorelasikan dengan penilaian kategori dosen. Diharapkan hasil penelitian ini dapat membantu pihak manajemen universitas dan dosen mengetahui hubungan antara nilai numerik dengan komentar mahasiswa. Kata Kunci—Text classification, Natural Language, Naïve Bayes, K-Nearest Neighbor, Support Vector Machine.