Mock Data Examples – Kor.ai

This page outlines sample inputs used for testing and validating the Bayesian risk models for insider dealing and spoofing. Each example includes:

• Raw business data
• Transformed node input (obfuscated format)
• Outcome explanation (used during model testing)

---

✅ Insider Dealing Case – Full Input Match

Raw Input

{
  "trade_size": 3_000_000,
  "instrument": "Brent Crude July",
  "news_sentiment": "Negative",
  "internal_chat": "Spoke to CEO prior to trade",
  "execution_timestamp": "2025-06-02T08:22:00Z",
  "employee_KYC_status": "Red flag",
  "access_to_sensitive_data": true
}
{
  "Q1": "High",
  "Q2": "Yes",
  "Q3": "Yes",
  "Q4": "Low",
  "Q7": "Yes",
  "Q8": "Yes"
}
{
  "trade_size": null,
  "news_sentiment": "Negative",
  "access_to_sensitive_data": false,
  "execution_timestamp": "2025-06-02T08:22:00Z"
}
{
  "Q1": "Unknown",
  "Q2": "Yes",
  "Q3": "Unknown",
  "Q8": "No"
}
{
  "quote_direction": "Buy",
  "quote_size": 2_000_000,
  "order_pulled_ms": 80,
  "execution_occurred": true,
  "side_of_execution": "Sell",
  "volatility_spike": "Yes"
}
{
  "S1": "High",
  "S2": "Yes",
  "S3": "Yes",
  "S5": "Yes"
}

---

Let me know once that’s in, and I’ll move you to Step 4: Agena API Integration (`03_Agena_API_Integration.md`). That one’s key for the inference side of your system.