Risk Detection in E-Commerce Using Markov Decision Process

The growth of e-commerce business especially in the business tocustomer and businessto -business relationships segments in recent years is an indication of how computer technology has improved activities in human lives. It is widely projected that ecommerce as a sector of business transaction is poised for spectacular growth. An immediate shortcoming of e-commerce is the risk involved in transactions itself. Such risks include viruses, hacking, repudiation, account takeover fraud, card generators, changing ship address, lost and stolen cards, delivery of physical goods, delivery of digital goods, legal and regulatory issues. These risks can be broadly classified into security, payment, fulfillment and legal and regulatory risks. Formal methods of risks assessment of e-commerce transactions have not being critically looked into in order to encourage this business method. We propose a model for assessment of e-commerce risks which is further simulated using available data. The system performance was further evaluated using MATLAB and is found to be more than 90% successful when compared with existing systems.