Thesis (M. Soc. Sc.)--University of Hong Kong, 1994. / Includes bibliographical references (leaves 107-112).

Thesis (M.Soc.Sc.)--University of Hong Kong, 1994. / Includes bibliographical references (leaves 107-112) Also available in print.

published_or_final_version / Criminology / Master / Master of Social Sciences

Thesis (M.B.A.)--Naval Postgraduate School, December 2003. / Thesis advisor(s): Donald Summers, Juliette Webb. Includes bibliographical references (p. 67-69). Also available online.

ID-theft survival kit -- Book From victim to victor -- ID theft FAQ -- Audiocassettes -- Identity theft resources -- Testimonials -- ID theft action letters -- About the author -- Media appearances -- Identity theft laws -- Theft Deterrence Act. / Title from opening screen, December 28, 1999.

The present four field studies explored the effectiveness of multiple prevention techniques designed to increase the frequency of cashiers' identification (ID)-checking behaviors from a customer-focused and cashier-focused approach. Studies 1 and 2 examined customer-focused approaches, whereas Study 3 examined a cashier-focused approach. Study 4 examined a combination of the cashier-focused and customer-focused approaches. From a customer approach, Study 1 investigated the use of four prompts (a no-prompt control, an antecedent only, an antecedent with a positive consequence, and an antecedent with a negative consequence) at encouraging cashiers to ask customers for their ID during a credit purchase. Research assistants (RAs) visited various stores and made credit purchases, while displaying one of the four prompts covering their card's signature line to the cashier during check-out. The results showed RAs were checked for ID the most when using the prompts containing the antecedent and consequence, which was checked for ID significantly more than the no-prompt control. Study 2 (also a customer approach) attempted to replicate Study 1 in a non-college community. Using a similar methodology as Study 1, the results showed RAs were checked for ID the most when using the prompt with the antecedent and positive consequence, which was checked for ID significantly more than the no-prompt control. From a cashier approach, Study 3 investigated the use of a goal-setting and prompt intervention led by the restaurant manager to increase the frequency of cashiers' ID-checking behavior. Using an A-B-A (Baseline-Intervention-Withdrawal) reversal design at one of two restaurants, the results showed the intervention restaurant's percentage of ID-checked purchases increased from Baseline to the Intervention phase. But, it decreased slightly during the Withdrawal phase, showing functional control but also some maintenance over the target behavior. The percentage of ID-checked purchases at the control restaurant was almost nonexistent throughout the study. Study 4 investigated the impact of using two intervention approaches (i.e., the customer and cashier approach) as opposed to one (i.e., the customer approach) to increase the frequency of cashiers' ID-checking behavior. While the A-B-A phases were occurring in the restaurants used in Study 3, RAs entered the restaurants and displayed an antecedent and positive consequence prompt to the cashiers during a credit purchase. The results of Study 4 partially supported the hypothesis. The cashiers in the intervention restaurant significantly checked more RAs for ID when two intervention approaches were combined than when only one intervention approach was used during Baseline, but not during the Withdrawal phase. / Ph. D.

Cashiers' identification-checking behaviors were observed at two grocery stores with the aim to actively involve cashiers in decreasing credit-card fraud. After baseline observations, cashiers at one store received a participative goal-setting and feedback intervention, whereby they collaboratively set a store goal for checking customers' identification. Over 23 days, the cashiers received one-to-one verbal feedback on their store's identification-checking percentages. The percentage of identification-checked purchases at the intervention store increased from 0.2 percent at Baseline to 9.7 percent during the Intervention. Then, it declined to 2.3 percent during Withdrawal, showing functional control of the intervention over the cashiers' target behavior. The cashiers at the other store served as the control group, and their percentage of identification-checked purchases were 0.3 percent, 0.4 percent, and 0.7 percent respectively during each of the A-B-A phases at the intervention store. It was also found the intervention affected male cashiers more than female cashiers. The present study also assessed the social validity of the current intervention by surveying both customers and cashiers from the intervention store. The results showed that customers do not mind getting their ID checked, while cashiers consider it important to check a customer for identification during a credit purchase. / Master of Science

Differential Evolution is an optimization technique of stochastic search for a population-based vector, which is powerful and efficient over a continuous space for solving differentiable and non-linear optimization problems. Weighted voting stacking ensemble method is an important technique that combines various classifier models. However, selecting the appropriate weights of classifier models for the correct classification of transactions is a problem. This research study is therefore aimed at exploring whether the Differential Evolution optimization method is a good approach for defining the weighting function. Manual and random selection of weights for voting credit card transactions has previously been carried out. However, a large number of fraudulent transactions were not detected by the classifier models. Which means that a technique to overcome the weaknesses of the classifier models is required. Thus, the problem of selecting the appropriate weights was viewed as the problem of weights optimization in this study. The dataset was downloaded from the Kaggle competition data repository. Various machine learning algorithms were used to weight vote a class of transaction. The differential evolution optimization techniques was used as a weighting function. In addition, the Synthetic Minority Oversampling Technique (SMOTE) and Safe Level Synthetic Minority Oversampling Technique (SL-SMOTE) oversampling algorithms were modified to preserve the definition of SMOTE while improving the performance. Result generated from this research study showed that the Differential Evolution Optimization method is a good weighting function, which can be adopted as a systematic weight function for weight voting stacking ensemble method of various classification methods. / School of Computing / M. Sc. (Computing)

Credit card fraud is a field with perpetrators performing illegal actions that may affect other individuals or companies negatively. For instance, a criminalcan steal credit card information from an account holder and then conduct fraudulent transactions. The activities are a potential contributory factor to how illegal organizations such as terrorists and drug traffickers support themselves financially. Within the machine learning area, there are several methods that possess the ability to detect credit card fraud transactions; supervised learning and unsupervised learning algorithms. This essay investigates the supervised approach, where two algorithms (Hellinger Distance Decision Tree (HDDT) and Random Forest) are evaluated on a real life dataset of 284,807 transactions. Under those circumstances, the main purpose is to develop a “well-functioning” model with a reasonable capacity to categorize transactions as fraudulent or legit. As the data is heavily unbalanced, reducing the false-positive rate is also an important part when conducting research in the chosen area. In conclusion, evaluated algorithms present a fairly similar outcome, where both models have the capability to distinguish the classes from each other. However, the Random Forest approach has a better performance than HDDT in all measures of interest.

In the past few years, Artificial Intelligence (AI) has evolved into a powerful tool applied in multi-disciplinary fields to resolve sophisticated problems. As AI becomes more powerful and ubiquitous, oftentimes the AI methods also become opaque, which might lead to trust issues for the users of the AI systems as well as fail to meet the legal requirements of AI transparency. In this report, the possibility of making a credit-card fraud detection support system explainable to users is investigated through a quantitative survey. A publicly available credit card dataset was used. Deep Learning and Random Forest were the two Machine Learning (ML) methodsimplemented and applied on the credit card fraud dataset, and the performance of their results was evaluated in terms of their accuracy, recall, sufficiency, and F1 score. After that, two explainable AI (XAI) methods - SHAP (Shapley Additive Explanations) and LIME (Local Interpretable Model-Agnostic Explanations) were implemented and applied to the results obtained from these two ML methods. Finally, the XAI results were evaluated through a quantitative survey. The results from the survey revealed that the XAI explanations can slightly increase the users' impression of the system's ability to reason and LIME had a slight advantage over SHAP in terms of explainability. Further investigation of visualizing data pre-processing and the training process is suggested to offer deep explanations for users.