<![CDATA[The Covid-19 pandemic that has hit Indonesia since the beginning of 2020 has had a major impact on the world of education, so that the learning process that was originally carried out face-to-face has turned into online learning. In such a situation, the University of Muhammadiyah East Kalimantan utilizes the Learning Management System (LMS) in an online learning system using the LMS Open Learning platform. The purpose of this study is to find the best attribute values using Correlation Based Featured Selection and to test the performance of the Naïve Bayes algorithm using Confusion Matrix. The attributes used after going through the feature selection are time spent on course, course completed, assignments, mid-semester exams and quizzes. The results of testing 178 data with a ratio of training data schemes and testing data of 70:30 produce an accuracy of 98.14%, 80:20 produces an accuracy of 97.22% and 90:10 produces an accuracy of 94.44%. Thus, the best accuracy is obtained at 70:30 data composition, which is 98.14%.Keywords: Accuracy level; Naive Bayes; Online learning; Prediction Abstrak. Pandemi Covid-19 yang melanda Indonesia sejak awal tahun 2020 memberikan dampak besar terhadap dunia pendidikan, sehingga proses pembelajaran yang semula dilakukan secara tatap muka berubah menjadi Pembelajaran Dalam Jaringan (daring). Dalam situasi seperti tersebut, Universitas Muhammadiyah Kalimantan Timur memanfaatan Learning Management System (LMS) dalam sistem pembelajaran daring menggunakan platform LMS Open Learning. Tujuan penelitian ini adalah mencari nilai atribut terbaik menggunakan Correlation Based Featured Selection dan menguji performa algoritma Naïve Bayes menggunakan Confusion Matrix. Atribut yang digunakan setelah melalui seleksi fitur adalah time spent on course, course completed, nilai penugasan, nilai Ujian Tenga Semester dan nilai quiz. Hasil pengujian 178 data dengan rasio skema data training dan data testing 70:30 menghasilkan akurasi sebesar 98,14%, 80:20 menghasilkan akurasi sebesar 97,22% dan 90:10 menghasilkan akurasi sebesar 94,44%. Dengan demikian, akurasi terbaik diperoleh pada komposisi data 70:30, yaitu sebesar 98,14%.Kata kunci: Akurasi; Naïve Bayes; Pembelajaran dalam jaringan; Prediksi]]>
Copyrights © 2022
