Sklearn Isolation Forest icon

Sklearn Isolation Forest

Isolation Forest for anomaly detection using scikit-learn

Actions2

Predict

AI-generated

Summary

Detect anomalies in incoming data using a pre‑trained Isolation Forest model.

Inputs

  • modelData (required) — JSON string containing the pickled Isolation Forest model (base64‑encoded).
  • predictFeatureColumns (required) — Comma‑separated list of feature column names to use for prediction.
  • pythonPath — Path to the Python executable with scikit‑learn installed (default: python3).

Output shape

An array of updated input objects; each object includes the original fields plus boolean is_anomaly and numeric anomaly_score; the first object also adds model, n_outliers, and n_inliers.

The operation never fails silently if a column is missing; missing columns raise an error unless continueOnFail is enabled. Output length equals input length.

Examples

Example 1: Predict anomalies on a table of sensor readings

modelData = {"model_pickle": "..."}; predictFeatureColumns = "sensor1,sensor2,sensor3"

Example 2: Batch prediction of a CSV‑derived workflow

modelData = {"model_pickle": "base64..."}; predictFeatureColumns = "featureA,featureB"

Example 3: Re‑use a trained model from a previous train operation

modelData = {"model_pickle": "base64..."}; predictFeatureColumns = "col1,col2,col3"

Discussion