Credit card fraud detection using AI and machine learning techniques is a vital application in the financial industry, aimed at safeguarding both customers and financial institutions from fraudulent activities. Enhancing model performance relies heavily on creating relevant features from raw data. For example, features such as transaction frequency, average transaction amount, and time since the last transaction can be calculated. Anomaly detection methods like Isolation Forests, One-Class SVM, and auto encoders help identify transactions that deviate from a cardholder's normal behavior. This detection task is typically framed as a binary classification problem, classifying transactions as either "fraudulent" or "legitimate." Common algorithms employed include Logistic Regression, Decision Trees, Random Forests, Support Vector Machines, and Neural Networks. While effective fraud detection mechanisms and methods like poly encryption and randomization are commonly used to secure transactions, they are not infallible.