Paper Title
Developing an Ai-Driven Fraud Detection System: A Machine Learning Approach

Abstract
The detection of financial fraud is critical for maintaining the safety and integrity of online transactions. To this end, this research articulates a holistic approach to the identification of fraudulentactivities by using machine learning algorithms, Logistic Regression, K- Nearest Neighbors, Decision Trees, Random Forest, and XGBoost classifiers. Our method begins with exploratory data analysis in an attempt to visualize trends about transaction transactions and identify significant features, followed by model train ingand evaluation on a well- structured dataset. We preprocess the data using feature engineering and standard scaling, and then compare multiple models based on their performance. Among the tested models, Random Forest proved to be the most accurate, making it a reliable solution for fraud detection. Additionally, we implemented a user input system that allows real-time fraud prediction based on specific transaction details. This study contributes to the development of automated fraud detection systems, helping financial institutions reduce risks and prevent losses. The implementation, done using Python libraries and documented in Jupyter Notebook, emphasizes simplicity and flexibility. Keywords - Detection of financial fraud, Machine learning, Random Forest (RF), Logistic Regression(LR), XGBoost (XGB), Decision Tree (DT), K-Nearest Neighbors (KNN), Feature engineering, Data visualization, fraud(f), non-fraud(n-f).