Abstract
The rapid rise of computing power causes an endless availability of raw data and pushes science to perform increasable analytics to draw conclusions about important information. This fast-changing pace of technology has revolutionized most parts of the business world and equally awares public and private sector to enhance the decision capability. Organizations are realizing the significant value of big data and applying large-scale analysis frameworks to improve business. By setting employees’ and customers’ attitudes as the core asset they tend to interpret successfully the analyses, enhancing their decision-making capacity.
In educational sector, the enormous increase of data amounts that higher institutions produce, hides a wealth of information that must unhidden through analytics environments as well. Especially in distance learning, through learning management systems (LMS) the sheer amount of data that is generated and the speed at which it is done so, leads higher education to adopt effective practices for managing large data’ production. Institutions’ success and global competitiveness are highly reliant on their ability to make use of big data and analytics meeting students’ demands for better services.
The aim of this dissertation is the development of a big data scale analysis framework to support customized and personalized learning environments. In each level of the educational process in distance learning, students and teachers log into LMS and create a big set of digital footprints that provide insights into their learning behavior. The big data scale analysis framework focus to identify trends and patterns of their engagements to provide a highly accurate picture of how engaged they are in the learning process in order to predict future outcomes. This environment enables stakeholders’ attention to focus on educational data for the provision of better and more timely feedback about students’ performance enhancing targeted support and personalized assistance. This work implements educational data mining techniques, learning analytics tools, social network analysis, and NoSQL systems as a big data scale analysis framework. It aspires to support stakeholders with accurate interventions tailored to circumstances and risks of students’ attendance in the courses. In such an ecosystem, stakeholders can be alerted in time for disengaged students and then to intervene for improving their retention over time. In addition, students could potentially become more active in the learning process, expanding the ultimate impact of learning. Dissertation’s framework is going beyond typical reports of students’ ability or performance. It focus on encouraging them for greater self-reflection in the course and minimizing the growing need of distance learning to draw reliable inferences for personalized learning, reducing the lack of physical presence.
Advisory Committee
Vasilios Verykios (supervisor), Professor, School of Science and Technology, HOU
Chris Panagiotakopoulos (member), Professor, Department of Primary Education, University of Patras
Dimitrios Kalles (member), Associate Professor, School of Science and Technology, HOU