Online Fraud Detection: A Decision Tree Approach

Abstract

In the digital age, online payment fraud has grown to be a serious issue, requiring effective detection methods to reduce monetary losses. This study introduces the Decision Tree Classifier algorithm, a machine learning-based method for fraud detection. The model learns patterns to categorise transactions as either genuine or fraudulent after being trained on a dataset that includes both legitimate and fraudulent transactions. The dataset is pre-processed to address unequal class distributions, and feature engineering techniques are used to improve model accuracy. Because of its interpretability and capacity to manage intricate decision boundaries, the Decision Tree Classifier was selected. Key performance indicators like accuracy, precision, recall, and F1-score are used to assess the model's performance. According to experimental results, the suggested method accurately and successfully separates fraudulent transactions. Furthermore, a comparison with alternative categorisation algorithms demonstrates how effective the Decision Tree model is at detecting fraud. The results show that fraud detection systems may be greatly improved by machine learning, which lowers false positives and increases financial security. This study offers a reliable and understandable model for real-time fraud detection, which supports ongoing efforts in financial cybersecurity. To boost performance, future studies might use deep learning algorithms or ensemble techniques to optimise the model.