Ongoing Research In Educational Data Mining
Department: Information and Communication Technology Education
Topic: Using K-Means Clustering Algorithm to Determine Learner Typologies for Project-based Learning
By Delali Kwasi Dake & Esther Gyimah
Background to the Study:
Data in higher education is increasing rapidly with less or no benefits to academic counsellors, students and management. Educational Data Mining (EDM) is a fast growing research area with advanced machine learning techniques to mine and better understand students learning behaviours and course re-design in academia (Baradwaj and Pal, 2012; Romero and Ventura, 2010). The ideal motive of using data mining techniques in EDM is to expose hidden data patterns which serve as a predictive tool in Education.
One fast course re-design trend that academicians are pursuing is the concept of project based learning by grouping students per study groups or for a project purpose. The idea behind this concept is to improve learner participation in group projects and help learners develop relevant skills (Perera, Kay, Koprinska, Yacef, and Zaïane, 2009). Using K-Means Clustering algorithm is a better data mining technique for grouping students if improving learners’ academic performance is of high relevance to academic authorities. This unsupervised algorithm can form the basis for grouping large data sets into number of clusters or groups.