Interpretation & discovery
Substructure analysis
Investigate internal variation and sublabels within an existing group.
Overview
Substructure analysis examines whether groups produced by an existing labeling run contain additional internal organization.
The resulting sublabels remain associated with the parent run rather than becoming disconnected analytical outputs.
Uses
- Inspect heterogeneity within an existing label
- Generate sublabels
- Retain parent-child analytical context
- Compare substructure with the original labeling result
Interpretation and limitations
- Substructure can reflect genuine heterogeneity or over-partitioning.
- The parent labeling result remains an important part of the interpretation.
- Additional structure should be evaluated rather than assumed to be meaningful.