Educational Big Data (EBD) and Learning Analytics (LA) have being attracted enormous attention in recent years. Data collection process is the first step of EBD and LA. Based on the data source, data collection can be classified into two categories: manual data collection, and automatic data collection. This paper describes two educational systems: SCROLL (System for Capturing, Reusing, Reminding Of Learning Logs) as manual data collection and, M2B (Moodle, Mahara, Booklooper) as automatic data collection.

View full abstract

This article describes the results of a study that reveals previously hidden facts from a large amount of behavioral data currently being collected in the educational field (educational big data) by individually excluding those factors that affect human behavior on the basis of cognitive psychological findings. Furthermore, actual cases are presented to show that the information revealed in this study can be a driving force for changing the learning behavior of individual children, and consequently solve various educational problems. In order to extract meaningful information from the big data collected from human behavior, a deep understanding of human behavior is essential. Conversely, with a deep understanding of humans, big data that are a simple mass of information could become an abundant source of information.

View full abstract

In the process of plastering an intermediate layer on an earth wall, the actions of the plaster craftsperson were focused on and the characteristics of this task were codified making tacit knowledge explicit. The codified data was then used to develop e-learning materials to support the acquisition of technical skills and this system’s effectiveness was subsequently evaluated. The e-learning materials were targeted at beginners with one or two years experience. Under the supervision of a currently working plaster craftsperson, the course included content on how to use the trowel, the work process, and aspects that are clarified through three-dimensional motion analysis, muscle activity analysis, and eye movement analysis. The beginner students were used to evaluate the e-learning materials. The areas evaluated were their posture when working and fatigue before and after work. The results showed a trend toward a lessening of mental fatigue through the use of the e-learning materials.

View full abstract

Institutional Research (IR) has been receiving much attention in Japanese higher education. In order to guarantee the educational quality of university, it has been discussed how to utilize the educational big data. In this paper, it is considered to construct models of students’ learning states using large-scale students’ learning data collected through the baccalaureate degree program based on some machine learning methods. In this research, data in 5 years are utilized in order to investigate the generalization ability of the models, and the performances of some machine learning methods are compared. From the experimental results, it is indicated that the models of students’ learning states with high generalization ability can be constructed. Its capability of application to enrollment management is also discussed from experimental results.

View full abstract