Automated network optimisation using data mining as support for economic decision systems

Rozaki, Eleni ORCID: https://orcid.org/0000-0002-1229-4013 2019. Automated network optimisation using data mining as support for economic decision systems. PhD Thesis, Cardiff University. Item availability restricted.

Preview	PDF - Accepted Post-Print Version Download (12MB) | Preview
	PDF (Cardiff University Electronic Publication Form) - Supplemental Material Restricted to Repository staff only Download (2MB)

Abstract

The evolution from wired voice communications to wireless and cloud computing services has led to the rapid growth of wireless communication companies attempting to meet consumer needs. While these companies have generally been able to achieve quality of service (QoS) high enough to meet most consumer demands, the recent growth in data hungry services in addition to wireless voice communication, has placed significant stress on the infrastructure and begun to translate into increased QoS issues. As a result, wireless providers are finding difficulty to meet demand and dealing with an overwhelming volume of mobile data. Many telecommunication service providers have turned to data analytics techniques to discover hidden insights for fraud detection, customer churn detection and credit risk analysis. However, most are illequipped to prioritise expansion decisions and optimise network faults and costs to ensure customer satisfaction and optimal profitability. The contribution of this thesis in the decision-making process is significant as it initially proposes a network optimisation scheme using data mining algorithms to develop a monitoring framework capable of troubleshooting network faults while optimising costs based on financial evaluations. All the data mining experiments contribute to the development of a super–framework that has been tested using real-data to demonstrate that data mining techniques play a crucial role in the prediction of network optimisation actions. Finally, the insights extracted from the super-framework demonstrate that machine learning mechanisms can draw out promising solutions for network optimisation decisions, customer segmentation, customers churn prediction and also in revenue management. The outputs of the thesis seek to help wireless providers to determine the QoS factors that should be addressed for an efficient network optimisation plan and also presents the academic contribution of this research.

Item Type:	Thesis (PhD)
Date Type:	Completion
Status:	Unpublished
Schools:	Computer Science & Informatics
Date of First Compliant Deposit:	19 November 2019
Last Modified:	04 Nov 2022 13:38
URI:	https://orca.cardiff.ac.uk/id/eprint/126926

Actions (repository staff only)

Edit Item