Anomaly Detection in Altair AI Studio: A Comparative Approach Using DBSCAN Clustering and Isolation Forest (PyOD)

The global banking industry is undergoing a significant digital transformation fueled by the widespread adoption of online and mobile banking services. While this evolution offers increased convenience and accessibility for consumers, it also exposes financial institutions to growing risks such as fraudulent transactions, irregular account activity, and compliance violations. Traditional rule-based systems and batch-processing approaches, which rely on predefined thresholds and historical fraud patterns, are increasingly inadequate in detecting subtle or previously unseen anomalies, especially in real time.

To address these challenges, we embarked on a project to identify hidden patterns, anomalies in financial transaction data, and practical applications of machine learning models within a low-code environment. This project employed unsupervised machine learning techniques—specifically density-based spatial clustering of applications with noise (DBSCAN) and Isolation Forest from the Python outlier detection (PyOD) library.