serin-ai commited on
Commit
e4188ea
·
verified ·
1 Parent(s): 4434b86

Create README.md

Browse files

# 📊 Cifer Fraud Detection Dataset

The **Cifer-Fraud-Detection-Dataset-AF** is a high-fidelity, fully synthetic dataset created to support the development and benchmarking of privacy-preserving, federated, and decentralized machine learning systems in financial fraud detection.

This dataset draws structural inspiration from the **PaySim simulator,** which was built using aggregated mobile money transaction data from a real financial provider operating in 14+ countries. Cifer extends this format by scaling it to **6 million samples,** optimizing for **federated learning environments,** and validating performance against real-world datasets.

> **Accuracy Benchmark:**
> Cifer-trained models on this dataset reach **99.93% accuracy,** benchmarked against real-world fraud datasets with **99.98% baseline accuracy**—providing high-fidelity behavior for secure, distributed ML research.

Files changed (1) hide show
  1. README.md +96 -0
README.md ADDED
@@ -0,0 +1,96 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ task_categories:
4
+ - tabular-classification
5
+ - feature-extraction
6
+ language:
7
+ - en
8
+ tags:
9
+ - fraud-detection
10
+ - finance
11
+ - federated-learning
12
+ - cifer
13
+ pretty_name: Cifer-Fraud-Detection-Dataset-AF
14
+ size_categories:
15
+ - 1M<n<10M
16
+ ---
17
+
18
+ # 📊 Cifer Fraud Detection Dataset
19
+
20
+ ## 🧠 Overview
21
+ The **Cifer-Fraud-Detection-Dataset-AF** is a high-fidelity, fully synthetic dataset created to support the development and benchmarking of privacy-preserving, federated, and decentralized machine learning systems in financial fraud detection.
22
+
23
+ This dataset draws structural inspiration from the **PaySim simulator,** which was built using aggregated mobile money transaction data from a real financial provider operating in 14+ countries. Cifer extends this format by scaling it to **6 million samples,** optimizing for **federated learning environments,** and validating performance against real-world datasets.
24
+
25
+ > ### Accuracy Benchmark:
26
+ > Cifer-trained models on this dataset reach **99.93% accuracy,** benchmarked against real-world fraud datasets with **99.98% baseline accuracy**—providing high-fidelity behavior for secure, distributed ML research.
27
+
28
+ ---
29
+
30
+ ## ⚙️ Generation Method
31
+ This dataset is **entirely synthetic** and was generated using **Cifer’s internal simulation engine,** trained to mimic patterns of financial behavior, agent dynamics, and fraud strategies typically observed in mobile money ecosystems.
32
+ - Based on the structure and simulation dynamics of PaySim
33
+ - Enhanced for multi-agent testing, federated partitioning, and async model training
34
+ - Includes realistic fraud flagging mechanisms and unbalanced label distributions
35
+
36
+ ---
37
+
38
+ # 🧩 Data Structure
39
+ | Column Name | Description |
40
+ |------------------|-----------------------------------------------------------------------------|
41
+ | `step` | Unit of time (1 step = 1 hour); simulation spans 30 days (744 steps total) |
42
+ | `type` | Transaction type: CASH-IN, CASH-OUT, DEBIT, PAYMENT, TRANSFER |
43
+ | `amount` | Transaction value in simulated currency |
44
+ | `nameOrig` | Anonymized ID of sender |
45
+ | `oldbalanceOrg` | Sender’s balance before transaction |
46
+ | `newbalanceOrig` | Sender’s balance after transaction |
47
+ | `nameDest` | Anonymized ID of recipient |
48
+ | `oldbalanceDest` | Recipient’s balance before transaction (if applicable) |
49
+ | `newbalanceDest` | Recipient’s balance after transaction (if applicable) |
50
+ | `isFraud` | Binary flag: 1 if transaction is fraudulent |
51
+ | `isFlaggedFraud` | 1 if transaction exceeds a flagged threshold (e.g. >200,000) |
52
+
53
+ ---
54
+
55
+ # 📁 File Organization
56
+ Total Rows: **6,000,000**
57
+ Split into 4 folders/files for large-scale and federated learning scenarios:
58
+ - `Cifer-Fraud-Detection-Dataset-AF-part-1-4.csv` → 1.5M rows
59
+ - `Cifer-Fraud-Detection-Dataset-AF-part-2-4.csv` → 1.5M rows
60
+ - `Cifer-Fraud-Detection-Dataset-AF-part-3-4.csv` → 1.5M rows
61
+ - `Cifer-Fraud-Detection-Dataset-AF-part-4-4.csv` → 1.5M rows
62
+
63
+ Format: `.csv` (optionally `.parquet` or `.json` upon request)
64
+
65
+ ---
66
+
67
+ # ✅ Key Features
68
+ - Fully synthetic and safe for public release
69
+ - Compatible with federated learning (cross-silo, async, or multi-agent)
70
+ - Ideal for privacy-preserving machine learning and robustness testing
71
+ - Benchmarkable against real-world fraud datasets
72
+ - Supports fairness evaluation via distribution-aware modeling
73
+
74
+ ---
75
+
76
+ # 🔬 Use Cases
77
+ - Fraud detection benchmarking in decentralized AI systems
78
+ - Federated learning simulation (training, evaluation, aggregation)
79
+ - Model bias mitigation and fairness testing
80
+ - Multi-agent coordination and adversarial fraud modeling
81
+
82
+ ---
83
+
84
+ # 📜 License
85
+ **Apache 2.0** — freely usable with attribution
86
+
87
+ ---
88
+
89
+ # 🧾 Attribution & Citation
90
+ This dataset was generated and extended by Cifer AI, building on structural principles introduced by:
91
+
92
+ **E. A. Lopez-Rojas, A. Elmir, and S. Axelsson** <br>
93
+ *PaySim: A financial mobile money simulator for fraud detection.* <br>
94
+ 28th European Modeling and Simulation Symposium – EMSS 2016
95
+
96
+ ---