This study aims to identify patterns in academic performance among students at the Faculty of Mathematics and Natural Sciences (FMIPA), Medan State University (UNIMED) using the K-Prototype clustering algorithm. The method was chosen due to its capability to handle mixed data types, combining both numerical and categorical variables. Data were collected through a structured questionnaire, and the sample size was determined using Slovin's formula, resulting in 373 student records for analysis. The clustering process revealed three distinct groups: students with stable academic records but low organizational engagement; highly motivated new students; and independent senior students active in extracurricular activities. Each cluster reflects unique combinations of GPA, credit load, learning behavior, and involvement. The findings provide a data-driven foundation for designing targeted academic advising and student support strategies.
Copyrights © 2025