Integrating Adaptive Assessment in CodeChum for Personalized Programming Education
Xavier David Maranga, Mark Vincent Pausanos, Stephine Sinoy, Cherry Lyn C. Sta. Romana, Catherine N. Arellano · 2024
This study examines the implementation of adaptive assessment in CodeChum, a platform for personalized programming education, using the Multilayer Perceptron (MLP) Machine Learning Algorithm. The algorithm's key feature is its ability to dynamically adjust the difficulty of programming problems to suit each student's skill level, demonstrating a notable accuracy of 97%. This approach aims to enhance the learning experience by providing tailored challenges that match individual competencies. The research involved a dataset of around 10,000 programming problems and included the participation of 81 computer science students from Cebu Institute of Technology-University. These adaptive assessments varied in difficulty based on real-time student performance metrics, intending to improve engagement and comprehension in programming. The findings suggest a significant boost in student interaction and understanding, underscoring the value of adaptive learning methodologies in education. Statistical analysis from the pre-test (M = 32.78, SD = 39.83) and post-test (M = 196.81, SD = 61.18) showed that there is a significant difference between the performance of the students before and after using the adaptive assessment, t (80) = -25.8, p < .001, two-tailed. This study contributes to educational technology by showcasing how machine learning can be effectively used to create personalized and engaging learning experiences. While the results are promising, they also highlight the necessity for further investigation into the long-term effects of adaptive assessment and its generalizability across different educational settings.