Cardiff University | Prifysgol Caerdydd ORCA
Online Research @ Cardiff 
WelshClear Cookie - decide language by browser settings

Knowledge-based approach to risk analysis in the customs domain

Loukakos, Panagiotis 2014. Knowledge-based approach to risk analysis in the customs domain. PhD Thesis, Cardiff University.
Item availability restricted.

PDF - Accepted Post-Print Version
Download (12MB) | Preview
[img] PDF - Supplemental Material
Restricted to Repository staff only

Download (1MB)


The aim of this PhD project is to develop a fuzzy knowledge-based approach in support of risk analysis in the Customs domain. Focusing upon risk management and risk analysis in the Customs domain, this thesis explores the relationship of risk with uncertainty, fuzziness, vagueness, and imprecise knowledge and it analyses state of the art detection techniques for fraud and risk. Special focus is given to fuzzy logic, ontological engineering, and semantic modelling considering aspects such as the importance of human knowledge and semantic knowledge in the context of risk analysis for the Customs domain. An approach is presented combining the fuzzy modelling and reasoning with semantic modelling and ontologies. Fuzzy modelling and reasoning is explored in the context of risk analysis and detection in order to examine approximate human reasoning based on human knowledge. Ontologies and semantic modelling are explored as an approach to represent domain knowledge and concepts. The purpose is to enable easier communication and understanding as well as interoperability. Risk management is broader, multi-dimensional process involving a number of task, activities, and practises. The presented approach is focused on examining the analysis and detection of the risk, based on the outputs of the risk management process with the use of ontologies and fuzzy rule-based reasoning. An ontological architecture is developed in the context of the presented approach. It is considered that such architecture is possible to enable modularity, maintainability, re-usability, and extensibility and can also be extended or integrated with other ontologies. In addition, examples are discussed to illustrate representation of concepts at various levels (generic or specific) and the modelling of various semantics. Furthermore, fuzzy modelling and reasoning are investigated. This investigation consists of literature research and the use of a generic research prototype (examination of Mamdani and Sugeno model types). From theoretical research, fuzzy logic enables the expression of human knowledge with linguistic terms and it could simulate human reasoning in the context of risk analysis and detection. In addition, Hierarchical Fuzzy Systems (HFS) or Hybrid Hierarchical Fuzzy Controllers (HHFC) approaches can be used to manage complexity especially for complex domains. Linguistic fuzzy modelling (LFM) is an aspect that should be considered during fuzzy modelling. From the generic research prototype, fuzzy modelling with the use of ontologies is demonstrated together with their integration in the context of fuzzy rule-based reasoning. It is also considered that Mamdani type of fuzzy models is easier to express human knowledge since the output can be expressed with linguistic terms. However, Sugeno type of fuzzy model could be used from adaptive techniques for optimisation purposes.

Item Type: Thesis (PhD)
Status: Unpublished
Schools: Engineering
Subjects: T Technology > TA Engineering (General). Civil engineering (General)
Uncontrolled Keywords: Fuzzy modelling; Knowledge representation; Fuzzy knowledge-based approach; Customs; Risk analysis
Date of First Compliant Deposit: 30 March 2016
Last Modified: 19 Mar 2016 23:40

Actions (repository staff only)

Edit Item Edit Item


Downloads per month over past year

View more statistics