Modeling User Profile in Gamification Based e-Learning Using k-Means Clustering

The current development of e-Learning is still in the form of simple e-Learning as a place to share materials and do the exercises given by lecturers. This should be a problem since such e-Learning activities are not different from conventional learning and do not facilitate a personalized learning process. Personalized learning is important to address students' different comprehension abilities in the learning process. Personalization can be done by characterizing student profiles. Therefore, we need a system to determine profiles by modeling user profiles in e-Learning. In this study, personalization was done by recognizing the users' characteristics obtained from internal factors including personality, achievement emotion, and student activity logs when engaging in the e-Learning process. Student activity logs were automatically recorded by the system. To find out the profile of existing user, grouping was done using the unsupervised machine learning method, namely k-Means Clustering. Two k-clusters were employed to determine the user profiles. Two grouped profiles indicated differences in characteristics of the user profile group. The results of the cluster evaluation show the value of Silhouette Score of 0.43 and Connectivity Score of 8.85.