Authors :
Anas Shehu; Alhassan Salihu; Abubakar Sani
Volume/Issue :
Volume 7 - 2022, Issue 6 - June
Google Scholar :
https://bit.ly/3IIfn9N
Scribd :
https://bit.ly/3PqhBOU
DOI :
https://doi.org/10.5281/zenodo.6827708
Abstract :
The research was conducted in privacypreserving data publishing, to our knowledge only a few
used educational datasets to address privacy and utility.
This research used sample questionnaires to investigate
the awareness of privacy and its application to student
records and also applying privacy to students’ datasets
of all tertiary institutions of Kebbi State, Nigeria.
Student datasets were obtained from Kebbi State
Polytechnic Dakin-gari which we used as a benchmark.
K-anonymity and l-diversity models were used with k
configurations and suppression limits of 10 and 50% in
the ARX 3.9.0 de-anonymization environment. The work
evaluates data privacy, quality, and execution time for
each k value and two variants of suppressions limit.
Experimental results demonstrate that the higher the
suppression the more balanced exists between privacy
and utility. It was observed that suppression of 50%
provides less anonymization time irrespective of k
compared to k values when suppression = 10%. This was
proved to be due to less time it takes anonymization to be
completed. Also, our work ranks six institutions from 1st
through 6th based on some parameters obtained via
questionnaire/responses on privacy threats. The work
however established that all students’ records are faced
with serious privacy threats as no institution employ any
privacy-preserving techniques. Consequently, the
research proposed a privacy framework for all six
schools to deploy for better preservation
Keywords :
Arx de-anonymization tool, Dakin-gari, kanonymity, privacy, quality, utility
The research was conducted in privacypreserving data publishing, to our knowledge only a few
used educational datasets to address privacy and utility.
This research used sample questionnaires to investigate
the awareness of privacy and its application to student
records and also applying privacy to students’ datasets
of all tertiary institutions of Kebbi State, Nigeria.
Student datasets were obtained from Kebbi State
Polytechnic Dakin-gari which we used as a benchmark.
K-anonymity and l-diversity models were used with k
configurations and suppression limits of 10 and 50% in
the ARX 3.9.0 de-anonymization environment. The work
evaluates data privacy, quality, and execution time for
each k value and two variants of suppressions limit.
Experimental results demonstrate that the higher the
suppression the more balanced exists between privacy
and utility. It was observed that suppression of 50%
provides less anonymization time irrespective of k
compared to k values when suppression = 10%. This was
proved to be due to less time it takes anonymization to be
completed. Also, our work ranks six institutions from 1st
through 6th based on some parameters obtained via
questionnaire/responses on privacy threats. The work
however established that all students’ records are faced
with serious privacy threats as no institution employ any
privacy-preserving techniques. Consequently, the
research proposed a privacy framework for all six
schools to deploy for better preservation
Keywords :
Arx de-anonymization tool, Dakin-gari, kanonymity, privacy, quality, utility
