As is well knowledge, one type of digital crime in which forensics are used is credit card fraud. We discuss an approach in the fraud detection interface region in this study. A convolutional network and imbalanced, highly skewed value-based data are the methods suggested for the identification of fakes. The dataset used in this instance is the highly skewed machine learning Kaggle dataset for credit card extortion detection. The evaluated highlights for both the fraud and non-fraud classes are 1. Examining extortion evidence was a crucial tool for maintaining financial divisions. These days, the least successful method for credit card theft detection is the counterfeit neural network. The current approach used to identify extortion is plagued by misclassifications and increasing false positive. In order to create a system that can recognize credit card theft with a high degree of accuracy in all circumstances, this research paper uses convolutional neural network architecture layers.