DC] Investigating student's learning processes in a Virtual Reality learning environment by analyzing their interaction behavior

Abstract

Educational researchers use physiological sensors and human ob-servers to collect data to conduct quantitative studies in VR Learn ing Environments (VRLE). The data collected using physiological sensors are used to understand the affective state of the learners. Whereas, human observers are the most widely used method to collect data to understand the behavior of the learners in a VRLE. However, the data collection using human observers is labor inten-sive and may get biased due to cognitive, social, and communicative causes. Moreover, the data provided by human observers need to satisfy the inter-rater reliability test in order to become valid. There has been little or no work done to automatically collect the data re-lated to the behavior of the learners in VRLE. To overcome the gaps mentioned above we have developed a data collection mechanism to automatically log all the interaction behavior of the learners happening in VRLE along with the time-stamp in real-time. The developed data collection mechanism can open up new opportunities for data collection when deployed in VRLEs. Also, the data collected using it can be useful to get more insights about the learners' behavior and in turn the learning processes happening in VRLE by applying AI and ML tools. This doctoral thesis aims at investigating the interaction behavior of the learners happening in VRLE and analyze the behavioral data to understand the behavioral pattern responsible for the higher learning gain. The analysis can also be further done for the early prediction of the learning outcomes based on the interaction behavior to provide scaffolds and adaptive feedback.