Cortex AI logo placeholderCortexLabAI · learning · engineering
Model Atlas · Anomaly detection

Isolation Forest

Detects anomalies by how quickly random trees isolate a point.

core mathematical viewanomaly score based on expected path length
Mental model

Understand it before memorizing it.

Unusual points are easier to separate using random splits.

Best fit

Where this model earns its place

Outlier detection

Start with a simpler baseline then compare this model using the same evaluation split and operational constraints.

Fraud screening

Start with a simpler baseline then compare this model using the same evaluation split and operational constraints.

Monitoring

Start with a simpler baseline then compare this model using the same evaluation split and operational constraints.

Strengths and limits

Trade-offs matter more than popularity.

Strengths

✓ Unsupervised

✓ Scales well

Limitations

△ Threshold interpretation

△ Can miss contextual anomalies

Evaluation

Metrics to watch

Precision@kInterpret with the product objective and error cost.
Recall@kInterpret with the product objective and error cost.
Alert rateInterpret with the product objective and error cost.
Production checklist

Before it reaches real users

  1. 01

    Tune contamination carefully

    Document the assumption and instrument the condition so regressions can be detected.

  2. 02

    Review alerts

    Document the assumption and instrument the condition so regressions can be detected.

  3. 03

    Track data drift

    Document the assumption and instrument the condition so regressions can be detected.