Data Analysis of Thesis Guidance Students Using Random Forest, Gradient Boosting, and Naïve Bayes Algorithms (Case Study: University of Raharja) Analisis Data Mahasiswa Bimbingan Skripsi Menggunakan Algoritma Random Forest, Gradient Boosting, dan Naïve Bayes (Studi Kasus: Universitas Raharja)

Abstract

Thesis guidance is a crucial stage in higher education, as the thesis is one of the primary requirements for earning a bachelor's degree. One of the main challenges in thesis guidance is managing consultation data between students and their supervisors. The application of technology and machine learning approaches offers significant potential in addressing this issue. Machine learning algorithms such as Random Forest, Gradient Boosting, and Naïve Bayes can be utilized to automatically analyze thesis guidance data, thereby assisting supervisors in efficiently monitoring student progress. This research aims not only to provide a solution for supervisors in monitoring the progress of their students but also to offer a valuable tool for university management to evaluate the performance of supervisors in providing guidance. Based on the results and comparisons conducted, it can be concluded that the Gradient Boosting method achieves the highest accuracy, reaching 100%, compared to Random Forest with an accuracy of 98.8% and Naïve Bayes with an accuracy of 97.4%. From the testing data results using the Naïve Bayes, Gradient Boosting, and Random Forest algorithms, different accuracy levels were observed. However, the prediction outcomes were consistent: out of 235 testing data, 25 data points were classified as "Not Eligible," and 210 data points were classified as "Eligible" based on the established criteria.