AI Ninja Program
Cyan eye
— Data & observation. See what the model sees.
Purple eye
— Intelligence & intuition. Think like a network.
Neural bridge
— Connecting data to decisions.
Hood & headband
— The ninja identity. Mastery through practice.
Orbiting nodes
— Data in motion. Always learning.
☀
☾
★
Bookmark this step
↻
Reset all progress
Clear all progress and bookmarks?
Yes, reset
Cancel
⚠
Are you sure?
Cancel
Confirm
←
Back to portal
↺
Model Evaluation
Metrics, cross-validation, and overfitting
0%
·
0 of 5 steps completed
·
~90 min · Learn why accuracy lies and what to use instead
1
The Accuracy Trap
Why 99% accuracy can mean zero detections
99%
2
Confusion Matrix
TP, TN, FP, FN — the four outcomes of every prediction
TP
FP
FN
TN
3
Precision, Recall, F1
The metrics that actually matter in security
P
R
F1
0.92
0.81
4
ROC Curve and AUC
Compare classifiers across all thresholds
5
Threshold Tuning
Choose the right operating point for your SOC