Please use this identifier to cite or link to this item: http://ir.futminna.edu.ng:8080/jspui/handle/123456789/18050
Title: An Intelligent Model for Detecting the Sensitivity Sentiments in Learners’ Profile on Learning Management System
Authors: Muhammad, M.K.
Ishaq, O. O.
Olayemi, M. O.
Ojeniyi, Joseph Adebayo
Keywords: Intelligent, Attributes, Sensitivity, Profile, Learner, Privacy, Security, LMS
Issue Date: 2021
Publisher: 3rdSchool of Physical Sciences Biennial International Conference
Series/Report no.: 3rdSchool of Physical Sciences Biennial International Conference;
Abstract: Mobile learning is enhanced through learner information analytics. Electronic learning systems are capable of offering personalized learning experiences with regards to learners’ distinct attributes including knowledge, skills, and competencies required in evolving effective learning and teaching. User profiles on the LMS networks is evolving area of research by utilising social networks plugins in order to harvest data for sentiment analysis. Recently, a high-level sentiment analysis is adopted for the purpose of understanding the opinions of learners concerning a specific product or trends from reviews or tweets. Therefore, sentiment analysis is useful in improving the understanding of learners/user opinion, and also extracting trends about privacy and security of profile information on the LMS. The automated schemes are trained with hidden patterns in the comments or reports, then assign diverse sentiments to different reports indicating the severity and sensitivity regarding privacy of attributes in the learner profile information supplied to the LMS during the registration. The existing approaches require manual prioritization and partitioning, which is complex and time-inefficient. The first phase is the design of an opened questionnaire to be adopted during voluntary interview sessions with selected learners on different LMS such as CODel, NOUN, and other LMSs for the purpose of generating learners profile information. In the second phase, 20 attributed were identified as most relevant for inclusion in mobile learner’s profile information in which (12) attributes were considered to be most-sensitive, more-sensitive (2), less-sensitive (3), nominal (3), and non-sensitive (0) respectively. In this regard, the paper proposes deep learning networks techniques such as convolutional neural network (CNN), LSTM- based for the purpose automatically discovering sensitive attributes requiring privacy and protection from the public access or exploitation. The outcomes were better when compared to the manual-based thematic approach previously utilised in terms of effectiveness and accuracy
URI: http://repository.futminna.edu.ng:8080/jspui/handle/123456789/18050
Appears in Collections:Cyber Security Science



Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.