Hierarchical Pivot Selection for Interpretable Data-Agnostic Models

This method uses proximity trees and ensemble learning to select representative pivot instances for constructing interpretable predictive models. The approach is data modality-agnostic and functions as both a selection technique and a standalone predictive model by comparing input instances to selected pivots.