metadata
license: apache-2.0
task_categories:
- tabular-classification
language:
- en
tags:
- finance
size_categories:
- n<1K
📊 Cifer Fraud Detection Mini Dataset
🧠 Overview
(cifer-fraud-detection-mini-dataset)
The Cifer-Fraud-Detection-Mini-Dataset is a lightweight sample containing 20 transaction records, extracted from the full 21 million-row Cifer-Fraud-Detection-Dataset-AF. It is designed for quick experimentation of encrypted model training with Fully Homomorphic Encryption (FHE).
Though small in size, this mini dataset retains the original schema and data structure inspired by PaySim—reflecting realistic financial transaction behaviors. It enables users to prototype and validate privacy-preserving ML workflows in environments where encrypted data pipelines and secure federated orchestration are required.