Credit Card Fraud Detection Web Application using Streamlit and Machine Learning

Credit Card Fraud is one of the major threads in the financial industry. Due to the covid-19 pandemic and the advance in technologies, the number of users is increasing, with the increased use of credit cards. Due to more use of credit cards, Fraud cases also increase day by day. The research community striving hard to explore myriad credit card fraud detection techniques, but changes in technology and the varying nature of credit card fraud make it difficult to develop an effective technique for the detection of credit card fraud. This research work used a real-world credit card dataset. To detect the fraud transaction within this dataset three machine learning algorithms are used (i.e. Random Forest, Logistic regression, and AdaBoost) and compared the machine learning algorithms based on their Accuracy and Mathews Correlation Coefficient (MCC) Score. In these three algorithms, the Random Forest Algorithm achieved the best Accuracy and MCC score. The Streamlit framework is used to create the machine learning web application.