An Improvement Of Matrix-Based Clustering Method For Grouping Learners In E-Learning

Kun Zhang,Li-zhen Cui,Haiyang Wang,Qi Sui

DOI: https://doi.org/10.1109/CSCWD.2007.4281577

2007-01-01

Abstract:Clustering learners into groups according to the customized features in e-learning environment is an important step to build a personalization learning system. Though clustering e-learners is important for better cooperation between teachers and students in e-learning, it is a challenge job to group learners flexibly and exactly. Since there are already many models for the features which are used for the basis of the clustering methods, this paper proposes an improvement of Matrix-based Clustering Method which preformed efficiently without extra comparison in contrast to k-means clustering algorithm. The improvement of the Matrix-based Clustering Method proposes the concept "Agglomerate Strength " for further cluster cohesion measurement in contrast to the previous Matrix-based Clustering Method in precision. And the comparison experiments between the improvement Matrix-based Clustering Method and the other methods, i.e. the previous Matrix-based Clustering Method and K-means algorithm, are investigated. The results of experiments show that this method is feasible and efficient.