Synthetic Outliers Evaluation

The task involves evaluating the quality and utility of tabular synthetic data generated with outliers in the context of machine learning. Outliers, generated according to Extreme Value Theory (EVT), help model rare events, which in turn enhances the realism of the generated data, improves prediction quality in machine learning, and facilitates the detection, handling, and removal of outliers in real-world data.