Data Mining Analysis of Student Perceptions of ChatGPT Usage Using the C4.5 Algorithm

Abstract

The rapid advancement of artificial intelligence, particularly ChatGPT, has brought considerable changes to the educational landscape. Despite its benefits, concerns have emerged regarding its potential negative impact on students' critical thinking skills and learning independence. This study aims to classify the perceptions of students at Dhyana Pura University toward the use of ChatGPT in education and to measure the accuracy of the resulting classification model. The method applied is data mining using the Knowledge Discovery in Databases (KDD) approach with the C4.5 algorithm. Data were collected from 350 active students through a Likert scale-based questionnaire covering five indicators: ease of use, perceived benefits, impact on critical thinking, impact on learning independence, and attitude toward ChatGPT usage. The classification process was carried out using Altair AI Studio with an 80:20 training and testing data split. The results indicate that the majority of students hold a neutral perception toward the use of ChatGPT in academic activities. The classification model achieved an accuracy rate of 86.96%, which is categorized as good. This research is expected to serve as a reference for educational institutions in formulating policies on the responsible use of AI in academic settings.