Penerapan Data mining Klasifikasi Tingkat Kepuasan Mahasiswa Terhadap Pelayanan Akademik Menggunakan Metode Naïve Bayes Dan Support Vector Machine (Studi Kasus Program Studi Sistem Informasi Universitas Labuhanbatu)

Abstract

This study was conducted to classify public satisfaction levels using the Support Vector Machine (SVM) algorithm as the primary data analysis method. The objective of this study was to obtain an accurate and reliable prediction model for determining the Satisfaction and Dissatisfaction categories based on the available data. The theoretical basis used refers to the concept of machine learning, specifically SVM, which works by forming an optimal hyperplane to separate data classes. In addition, model evaluation theories such as the Confusion Matrix were used to objectively measure prediction performance. The research methodology included data collection, pre-processing, dividing the dataset into training and test data, and training the SVM model. Evaluation was conducted using accuracy, sensitivity, and specificity metrics to assess the model's ability to predict data accurately. The results and discussion indicate that the SVM successfully classified the majority of data correctly, with the Satisfaction class having a perfect prediction rate while the Dissatisfaction class still had a small error. Further analysis indicated the need for SVM parameter optimization to improve accuracy in the minority class. The conclusion of this study states that the SVM has good performance in classifying public satisfaction data, although it still requires refinement in recognizing certain class patterns. This finding opens up opportunities for developing more adaptive methods to improve predictive performance.