Analisis Dan Identifikasi Pola Perilaku Penggunaan Aplikasi Microsoft Copilot Oleh Guru Sma Di Jawa Tengah Menggunakan Pendekatan K-Mean Clustering

Abstract

Digital transformation in education has become an important need in today's technological era. The development of artificial intelligence (AI)-based applications, such as Microsoft Copilot, offers great potential to improve the efficiency and effectiveness of the learning and administration process. This application can help teachers in compiling lesson materials, analyzing student data, and automating various administrative tasks. This study was conducted on high school teachers in Central Java whose technology implementation was already running but their behavioral patterns of use had not been properly mapped. The purpose of this study was to analyze and identify behavioral patterns of Microsoft Copilot application use by high school teachers in Central Java using the K-Mean Clustering approach. The analysis method in this study uses the K-Mean Clustering approach, which is one of the analysis methods that can be used to group data into several clusters based on certain characteristics. K-Mean Clustering is applied to identify behavioral patterns of Microsoft Copilot use by high school teachers in Central Java. As a data processing tool, RapidMiner 12.1 software is used The results of this study indicate that the application of k-means clustering based on 19 parameters measured using a questionnaire regarding the application of Microsoft Copilot to 238 samples of high school teachers in Central Java obtained 2 optimal clusters, where the k-means clustering experiment using k = 2, k = 3, k-4 and k-5 showed that the analysis results for k = 2 had the lowest Favies Bouldin Index (BDI) value. The results showed that the classification with k = 2 had the highest validity. The classification results obtained that cluster 1 which is a cluster with a score category is only occupied by 1 teacher, while cluster 0 which is a high value cluster is accompanied by 237 other teachers. Suggestions for further research are to apply classification/cluster analysis using different methods, for example Hierarchical Clustering. In addition, by observing previous data, 1 sample (respondent) of teachers seems to have an extreme value compared to other scores. Therefore, the suggestion for further research is to exclude one data..