| المستخلص: |
Microblogging– a kind of blogging where users publish snippets of information about their daily activities and thoughts over the Internet– has quickly become popular during the last few years. On Twitter as an example of microblogs, millions of people post short text based updates– tweets– about broad range of issues. Topics range from their personal life and work, to current events, news, and interesting observations and political thoughts. In the midst of this general acceleration in using social media, there is an increase in using it in learning, in particular microblogging environments such as Twitter. As a result, there is an increase in the amount of data within these social media that reflects what people are learning and how well they are learning. For example, courses that adopt Twitter in the learning process allow students to discuss with each others and with their teacher different topics and to express their opinions on various aspects of these topics. The data generated out of these discussions constitutes a valuable resource on which teachers can rely for courses’ or learning activities evaluation. To understand this complicated landscape of topics that people are learning during the life time of the learning process, and how well these topics are being learnt, it would be helpful if they represented into a network . This network organizes topics being discussed in these microblogs according to their various subtopics and stored information about expressed opinions on these topics. This envisioned network would provide information that would help to track the learning process over time and to observe the strengths and the limitations in the learning process while it is happening. In this research, we aim to study the possibility of mapping microblogs generated in learning activities into a network of topics by utilizing methods from Microblogs Analytic, Learning Analytic and Text Mining. This network would provide essential feedback about what topics are being learnt by people, the size of interest in particular topics and how well these topics are being learnt. We conduct several experiments to develop and evaluate the framework that maps the microblogs generated into a network of topics. End to End evaluation of the complete system achieved 87% accuracy based on the used measurements.
|
