Mkuzangwe, Nenekazi Nokuthala Penelope (DPhil)
Nenekazi Mkuzangwe holds a Master's degree in Science in Mathematical Statistics, cum laude , from Rhodes University. She was awarded a Rhodes Half Honours Scholarship for passing her third-year Mathematical Statistics with distinction. She has worked at the Nelson Mandela University where she lectured Statistics to Science, Health Science and Commerce students. She has mentored university students in applying Statistics-based machine learning techniques to analyse real life data in a project called Data Science for Impact and Decision Enablement sponsored by the Department of Science and Innovation. She is currently employed by the CSIR as network and data security researcher. Intrusion detection is the process of identifying whether an unauthorised access on or unauthorised attempt to access an information system is occurring or has occurred. Intrusion detection systems have been proposed to perform intrusion detection in information systems; however, there is no frame of reference to measure classification accuracies of these systems. The research conducted by the candidate addressed this problem by empirically determining the achievable upper bounds on the classification accuracies of two ensembles of classifiers based network intrusion detection systems. These bounds are the first to be defined in terms of information gain and data entropy. The research also addressed the unavailability of real world network trace due to privacy and legal restrictions by applying differential privacy to preserve the privacy of the number of TCP synchronisation packets associated with HTTP requests. The work has been published in an international journal and conferences held in China and Japan.
Supervisor: Prof FV Nelwamondo
123
Made with FlippingBook - Online Brochure Maker