The online transaction is one of the digital technologies. The number of credit card transactions, including fraudulent ones, is rising daily. The credit card fraud detection system is used to find fraudulent transactions in the financial sector. To get around this problem, researchers used a few machine learning methods. The experimental research in this paper was conducted using the European Credit Card Fraud (CCF) dataset. The dataset was divided into a train set (60%) and a test set (40%). The following algorithms were used in the experiment: Random Forest, XGBoost, AdaBoost, Decision Tree, and Logistic Regression, CNN. The results show that logistic regression produces more accurate and valuable results utilizing assessment measures than Random Forest, XGBoost, AdaBoost, and Decision Tree. In this paper, credit card transaction fraud is examined and the effects of data imbalance on model performance are determined using machine learning techniques. The models are evaluated using matrix metrics including accuracy, false positive rate, true positive rate, and confusion matrix.