Welcome to FedImpute

FedImpute is an open-source framework implemented in Python for distributed imputation research for horizontally partitioned data.

Features of FedImpute

• Flexible Distributed Missing Data Scenario Builder: Provide flexible API to construct distributedmissing data scenarios under various missing data distribution and data partition strategies or naturally distributed data with existing missingness.
• Built-in Distributed Imputation Methods: Supports multiple distributed imputation techniques, including mean, and model-based approaches, tailored for distributed data.
• Easy Integration: Designed to be easily extended with distributed imputation algorithms and workflows.
• Customizability: Offers extensive configuration options to adapt the imputation process to specific needs.

Quick Start

• Overview

• Installation and Basic Usage

User Guide

• Data Preparation
• Distributed Missing Data Scenario Builder
• Distributed Imputation
• Evaluation

Tutorials

Support and Contact

FedImpute is developed by Rutgers Institute for Data Science, Learning, and Applications (i-DLSA), lead by Professor Jaideep Vaidya. For any questions, please contact Sitao Min for support.