ANOMALY DETECTION IN ELECTRONIC BUSINESS

ABSTRACT

Outlier detection is the process of finding data objects with behaviors that are very different from expectation. In one of the most prominent use cases of anomaly detection, it is common to hear about events where one’s credit card number and related information get compromised. The closing of physical bank branches (due to COVID-19 pandemic), and the convenience of digital channels have all led to a rapid shift to online banking services. While it is generally seen as a positive outcome, it has caused particular security problems for customers who have never had to use their computers or smartphones before in order to manage their money. Often these people are not technologically savvy enough to recognize social engineering patterns, which aim to persuade a person to perform an action that they otherwise wouldn’t do. Since It is much harder for banks to accurately authenticate a customer they have never met and fraudsters are finding a higher rate of success in submitting fake documentation or stolen personal information for electronic business. In this study, a Machine Learning classifier algorithm was developed to minimize the problem posed by anomaly in electronic business. The application was developed with anocanda spyder 3. The program used datasets obtained from e-commerce stakeholders on condition of unanimity for training the Machine Learning application. Test run shows a high level of success in detecting anomaly in electronic business.