Classification of junior high school students' mathematics learning profiles using fisher's discriminant analysis

Abstract

The diverse characteristics of students in understanding mathematical concepts often hinder learning effectiveness if not accurately mapped. This study aims to classify students' mathematical learning profiles into analytical, collaborator, theoretical, and problem-solver types, while identifying the dominant factors that differentiate these groups. A descriptive quantitative approach was employed for this research. The population consisted of students at a junior high school in Banda Aceh, with a sample of 100 students selected through purposive sampling. Data were collected via questionnaires and subsequently analyzed using Fisher’s Discriminant Analysis with the enter method to develop an accurate classification model. Key findings revealed that all predictor variables collectively made a significant contribution, with learning styles and motivation emerging as the most dominant predictors in differentiating the four groups. Learning style acts as the fundamental basis for information processing, while environmental conditions serve as a strong characteristic for the collaborator group. The resulting discriminant function model demonstrated high reliability, achieving a classification accuracy of 96% for the original data and 89% in the cross-validation test. These results prove the reliability of the discriminant function in precisely mapping learning profiles based on the internal and situational factors perceived by students. These findings provide a strategic reference for educators in designing differentiated learning strategies that are more relevant to students' individual needs.