Intelligent agent service for detecting impersonators in online examination environment using K-nearest neighbor alorithm
| dc.contributor.author | Izu-Okpara, Ijeoma Onyinye | |
| dc.date.accessioned | 2025-04-11T11:03:19Z | |
| dc.date.available | 2025-04-11T11:03:19Z | |
| dc.date.issued | 2020-10 | |
| dc.description | A Master's thesis on detecting impersonators in online examination environment | |
| dc.description.abstract | Intelligent Agent Service for Detecting Impersonators in Online Examination Environment was developed for managing major challenges such as security and cheating (impersonation) that is now a critical issue associated with online examination system or computer-based testing (CBT). The key informant interview technique, observations, and critical review of articles related to CBT methodologies (design and development) were used to gather facts regarding the study area. The agile software methodology was adopted as the software development life cycle based on its strengths in team work and efficient product delivery haven examined six different software methodologies. A multi-level security service was developed to handle various security threats at the different operational levels of the proposed system using 256-bit SSH algorithm, Merssene Twister Algorithm, 128-bit Advanced Encryption Standard (AES), and Message Digest (MD) 5 algorithm. The K-Nearest Neighbor (KNN) machine learning algorithm was implemented as an intelligent agent service to give the proposed system some intelligence in detecting and classifying a likely suspected case of impersonation and its severity level during an online examination. JavaScript, Hypertext preprocessor (PHP), MySQL, Hypertext Markup Language (HTML), Cascading Style Sheet (CSS) and Python programming language were used to develop the software prototype. Unified Modeling Language (UML) such as sequence diagram and usecase diagram were used to model the system behavior and interactivity. The developed system accuracy in terms of the algorithm used for detecting impersonators was evaluated using confusion matrix. Results revealed 98% accuracy with the K-NN algorithm implemented. Finally, the results based on the acceptance testing done also revealed that 86% of the users strongly agreed with the performance level of the developed platform; hence, recommendations is made for the proposed system to be adopted by Nigerian Universities and companies based on its effectiveness in impersonation detection. | |
| dc.identifier.citation | Izu-Okpara, I. O. (2020). Intelligent gent service for detecting impersonators in online examination environment using K-nearest neighbor alorithm. (Unpublished Master's Thesis). Federal University of Technology, Owerri. | |
| dc.identifier.uri | https://repository.futo.edu.ng/handle/20.500.14562/1761 | |
| dc.language.iso | en | |
| dc.publisher | Federal University of Technology, Owerri | |
| dc.rights | Attribution-NonCommercial-NoDerivatives 4.0 International | en |
| dc.rights.uri | http://creativecommons.org/licenses/by-nc-nd/4.0/ | |
| dc.subject | K-NN algorithm | |
| dc.subject | classification | |
| dc.subject | intelligence | |
| dc.subject | online examination | |
| dc.subject | agile model | |
| dc.subject | Department of information management technology | |
| dc.title | Intelligent agent service for detecting impersonators in online examination environment using K-nearest neighbor alorithm | |
| dc.type | Master’s Thesis |