Electronic capture and analysis of fraudulent behavioral patterns : an application to identity fraud
Document Type
Dissertation
Date of Award
Summer 8-31-2005
Degree Name
Doctor of Philosophy in Information Systems - (Ph.D.)
Department
Information Systems
First Advisor
Michael Recce
Second Advisor
Marilyn M. Tremaine
Third Advisor
George Robert Widmeyer
Fourth Advisor
David Mendonca
Fifth Advisor
Joseph Wilder
Abstract
The objective of this research was to find a transparent and secure solution for mitigating identity fraud and to find the critical factors that determine the solution's acceptance. Identity fraud is identified as a key problem with total losses exceeding fifty two billion dollars (Javelin Strategy and Research 2005). A common denominator in most identity-fraud-prone transactions is the use of a keypad; hence this research focuses on keypad data entry and proposes a biometric solution. Three studies develop, evaluate and investigate the feasibility of this solution.
The first study was done in three stages. Stage one investigated the technical feasibility of the biometric keypad, stage two evaluated the keypad under different field conditions and stage three investigated acceptable user parameters. A key shortcoming with current authentication methods is the use of external identifiers that are prone to theft, unlike biometric patterns. A biometric keypad that supplements the present external identifiers was proposed, prototyped and evaluated. The results demonstrated that a biometric keypad can be a feasible medium performance solution. Addition of pressure and higher typing speeds were found to enhance discrimination accuracy while typing patterns were found to vary with elapsed time which led to deterioration in accuracy. The second study interviewed executives with experience in the introduction of new technologies with the objective of identifying and ranking critical factors that are important in the adoption of new biometrics. Performance, ease-of-use and trust-privacy issues were the most cited factors. A biometric acceptance model was formulated and five hypotheses were proposed from these interviews and prior research. Executives rated the keypad's ease-of-use high in comparison to other biometric approaches but were concerned about its accuracy.
The third study was a user attitude survey whose objective was to validate the formulated biometric acceptance model and acquire data on acceptable usage parameters. The proposed biometric model was validated and the proposed hypotheses were supported. Acceptable error rates and training times indicated that the biometric keypad would be more complex to engineer.
The dissertation concludes by summarizing the contributions and limitations of the three studies followed by several suggestions for future research.
Recommended Citation
Ngugi, Benjamin, "Electronic capture and analysis of fraudulent behavioral patterns : an application to identity fraud" (2005). Dissertations. 732.
https://digitalcommons.njit.edu/dissertations/732
