Learning Management Systems (LMS) is used widely around the world and its implementation is boosted due to the Covid-19 pandemic. With the rapid evolving research discipline of learning analytics and the vast data generated on LMS, this is useful to optimize the implementation of LMS and improve learning outcomes. Therefore, the research questions of this study are 1) What is the learners' online learning behavior on LMS? 2) What is the relationship between learners' online learning behavior and performance on LMS? And 3) Which learners' online learning behavior predict their performance on LMS? In this study, learners' online learning behavior on LMS refers to login time on LMS, number of downloads, and interaction with peers. On the other hand, learners' performance is their time spent in the system, number of exercise that students performed, number of forums posted, and their final course score. Data was collected from a LMS which involved 122 students. The results revealed that interaction with peers (behavior) predicted the number of forum posts (performance) significantly and the number of downloads (behavior) and interaction with peers (behavior) significantly predicted number of exercises performed (performance). Lastly, the total number of logins (behavior) also predicted time spent in the system (days) significantly. In conclusion, the findings support the potential for early prediction of learners' performance based on their behavior on LMS and demonstrate the benefits of identifying significant indicators from LMS data to facilitate successful online learning.

Learning Management Systems, Learning Analytics, Online Learning