================================================================================
ZERO-DAY ATTACK DETECTION EVALUATION REPORT
================================================================================
Generated: 2026-01-30 09:11:33
Evaluation Runs Analyzed: 37
================================================================================

SUMMARY STATISTICS:
----------------------------------------
Total Zero-Day Samples: 5933
Zero-Day Attacks Detected: 895
Zero-Day Attacks Missed: 4341
Zero-Day False Alarms: 246

Zero-Day Detection Rate: 15.09%
Zero-Day Precision: 78.44%
Zero-Day Recall: 17.09%
Zero-Day F1-Score: 28.07%

CONFUSION MATRIX ANALYSIS (Zero-Day vs Non-Zero-Day):
----------------------------------------
True Positives (Zero-Day ? Zero-Day): 895
False Positives (Non-Zero-Day ? Zero-Day): 246
False Negatives (Zero-Day ? Non-Zero-Day): 4341

STATISTICAL SIGNIFICANCE ANALYSIS:
----------------------------------------
Mean Detection Rate: 39.52%
Standard Deviation: 0.4052
95% Confidence Interval: [26.47%, 52.58%]

PERFORMANCE EVALUATION:
----------------------------------------
Overall Rating: VERY POOR
F1-Score: 28.07%
Recommendation: Zero-day detection is ineffective. Major architectural changes needed.

DETAILED ANALYSIS:
----------------------------------------
False Alarm Rate: 21.56%
Missed Detection Rate: 82.91%

================================================================================
END OF ZERO-DAY EVALUATION REPORT
================================================================================