<![CDATA[The growing complexity of learning activities in higher education highlights the need for accurate and scalable mechanisms to identify students’ learning preferences. Conventional VARK-based assessments, which rely on manual self-report questionnaires, remain limited by subjectivity, low practicality, and the absence of real-time feedback. This study addresses these challenges by developing Bandela, a mobile intelligent system that integrates the VARK framework with the K-Nearest Neighbors (KNN) classification algorithm to provide automated learning style identification. Using a Research and Development (R&D) approach, the system was implemented through a three-tier architecture consisting of a Flutter frontend, a Python Flask backend, and a MySQL database. Questionnaire responses collected from students were used as both training and testing datasets for the KNN model, enabling real-time classification across Visual, Auditory, Read/Write, and Kinesthetic categories. Functional evaluation through Blackbox Testing demonstrated that all core features ranging from authentication and questionnaire completion to classification processing, visualization, and community interaction performed reliably and as intended. The findings indicate that Bandela offers an accessible and empirically grounded tool for identifying learning preferences, contributing to more personalized and adaptive learning strategies. This work underscores the practical value of mobile intelligent systems in advancing data-driven personalization within higher education and provides a foundation for future enhancements involving expanded datasets and exploration of additional machine learning techniques.]]>
