In the past decades, ubiquitous learning (u-learning) has been the focus of attention in educational research across the world. Majority of u-learning systems have been constructed using ubiquitous technologies such as RFID tags and cards, wireless communication, mobile phones, and wearable computers. There is also a growing recognition that it can be improved by utilizing ubiquitous learning logs collected by the u-learning system to enhance and increase the interactions among a learner, contexts, and context-based knowledge. One of the issues of analytics based on u-learning is how to detect or mine learning logs collected by u-learning systems. Moreover, it is necessary to evaluate whether the recommendations detected by analysis are appropriate in terms of learning levels, contexts and learners' preference. To tackle the issues, we developed a system that could recommend useful learning logs at the right place in the right time in accordance with personalization of learners. An experiment was conducted to evaluate the system's performance and the recommendations' usefulness for learning. In the evaluation experiment, we found important criteria for recommending in the real-world language learning. In addition, the participants were able to increase their learning opportunities by our recommendation method.

ubiquitous learning, context-aware learning, spatio-temporal data mining, association analysis, ubiquitous learning analytics