TabEval Documentation#
Welcome to TabEval, a comprehensive Python framework for evaluating synthetic tabular data generation methods.
Overview#
TabEval provides a unified interface for benchmarking various synthetic data generation algorithms across multiple evaluation dimensions including statistical fidelity, utility preservation, privacy protection, and structural consistency.
Key Features#
Comprehensive Evaluation Metrics: 50+ evaluation metrics across multiple dimensions
Rich Plugin Ecosystem: Support for 15+ state-of-the-art synthetic data generation methods
Flexible Benchmarking: Easy-to-use benchmarking suite with caching and reproducibility features
Multi-Domain Support: Tabular data, survival analysis, time series, images, and domain adaptation
Extensible Architecture: Plugin-based design for easy integration of new methods and metrics
Quick Start#
Installation#
pip install tabeval
Table of Contents#
User Guide