Configuration
Adjust InsideForest for different dataset sizes and business goals.
Automatic presets
auto_fast: limits forest depth and size for quicker exploration.auto_feature_reduce: removes redundant variables before clustering.
Manual parameters
| Parameter | Role | Recommendation |
|---|---|---|
n_sample_multiplier |
Samples per tree. | Use 0.05 for quick tests, up to 0.15 for more detail. |
ef_sample_multiplier |
Exploratory sampling. | Increase to 15 for imbalanced datasets. |
method |
Cluster consolidation. | menu for precision, select_clusters for speed. |
Execution environments
- Local: use
venvorcondaand runpytest -q. - Server: enable logging and persistence with
.save(). - PySpark: adjust partitions with
spark.sql.shuffle.partitions.
Continue with Interpretation to learn how to present results.