Deceptive Plastic: Unveiling the World of Financial Fraud

As human society is developing, so is electronic commerce and with that the use of credit card are increasing. And along with this growth, as people prefer to shop using credit cards in both mode of payments, the possibility of fraud and fraud of credit card is also on rise. The criminals may steal the card physically, but more frequently, they use digital ways to obtain the credit card details along with the personal information linked to it in order to engage in illicit actions. There is a lot of overlap between credit card fraud and identity theft. Indeed, credit card theft is one of the most common forms of identity theft. In these cases, the personal information of victims is stolen to somehow get access to cards by registering a new account without the owner’s knowledge and certainly it is a form of cyberattack as well as data breach and is offensive crime. This behavior is both identity fraud and credit card fraud. In this study, we employ a CAT Boost, ADA Boost, LIGHT GRADIENT BOOSTING and CONVOLUTIONAL NEURAL NETWORK model to simulate the process of credit card transaction processing and display how it can be applied to the spot of such frauds. The dataset inspection method, factor determination, and techniques used for detection have a significant influence on the demonstration of fraud detection in credit card transactions. The accuracy for models can be determined by finding precision, recall, f1score, accuracy values but to compare the models, we will take ROC AUC Score into account. So, the ROC AUC Score of CAT BOOST, ADA BOOST, LIGHT GRADIENT BOOSTING and CONVOLUTIONAL NEURAL NETWORK model are 92.44, 79.74, 93.34, 95.25 respectively. And CONVOLUTIONAL NEURAL NETWORK is best among four with 95.25 ROC Score.