End to End Project - Financial Risk Profiling via Anomaly Detection

The objective of this project is to apply Unsupervised Machine Learning to detect fraudulent activity within a financial dataset. We will move beyond simple rule-based filters to implement advanced detection techniques that isolate high-risk anomalies while ensuring a seamless experience for legitimate users.

Learning Outcomes:

• Exploratory Data Analysis (EDA): Learning how to establish a "normal" baseline in a dataset using statistical distributions and measures of spread.
• Anomaly Detection: Understanding how to use Isolation Forest algorithms to automatically detect outliers in data without using pre-existing labels.
• Correlation Analysis: Developing the ability to identify relationships between multiple variables to uncover hidden patterns that a single variable might miss.