Interpretation

Translate InsideForest results into clear narratives for technical and business teams.

Checklist

• Present an executive summary (number of clusters, key metrics).
• Show standout rules using tables or bullet points.
• Include visualizations created with plot_multi_dims or feature-importance charts.
• Connect findings to concrete actions (priority segments, target customers).

Narrative example

Cluster A groups young customers with high online spending; it represents 18% of the base and converts 1.4× above average.

Exporting results

• explain_regions() and regions_: canonical region tables with geometry and supervised metrics.
• assign_regions(): row-level evidence, including overlap counts and explicit unmatched rows.
• generate_descriptions(...): produces short text snippets for reports.
• MetaExtractor: documents parameters used and feature masks from your catalog.

Describe region_target_class as enrichment metadata, not as a predicted class. Always disclose coverage and unmatched rate next to purity or lift.

Pair this guide with the FAQ and benchmarks to present quantitative evidence.