Assessing how Learner attributes affect learning preference using Explorative Big Data Analysis

Abstract

This study explores how different learner attributes affect learning preferences among learners. The study investigates the relationship between attributes such as gender, age, study duration and educational level, against learner preferred learning methods. The study applied an analytical cross-sectional approach using explorative big data machine learning models on a sample of over 350 learners against 19 variables. The study was conducted in 4 learning institutions within the Chikuni Mission of Monze district in the Republic of Zambia. The study's findings indicate that more female learners preferred to learn in cooperative groups with a steadily increasing interest in project learning methods. In contrast, males in primary schools preferred to learn independently. Project methods became popular among males at the secondary school level. The Random Forest Classifier model on a 0.32 train test data split showed a target prediction accuracy of 73%. While age, level of learning, hours of study and the visual domain were fundamental to predicting learning preference, attributes such as tribe, number of siblings, and auditory and kinesthetic domains seemed to contribute less significantly to the target variable. This paper submits that big data and machine learning explorative methods are principal to predicting the learning preferences of learners in classrooms. The study further postulates that creating preferred learning opportunities for learners has great potential to positively affect learning outcomes and create active and performing learners in our schools.