Taxonomy

In the data context, a taxonomy is a hierarchical classification system. It defines which categories exist, how they relate to each other, and which properties apply per category. Examples are biological taxonomies (class, order, family, genus, species), product taxonomies (eCl@ss, ETIM, UNSPSC), or content taxonomies for search indexes.

For AI and ML applications, the taxonomy is the structural foundation. A classification model can only sort into categories that the taxonomy defines. The granularity of the taxonomy determines the granularity of predictions. Too coarse a taxonomy loses discriminating power, too fine becomes hard to train because too few training examples exist per category.

Pimcore PIM manages taxonomies as a structural data model element. Multiple parallel taxonomies are natively supported (eCl@ss, ETIM, custom structures), AI-driven automatic classification can sort products against any maintained taxonomy, workflows let human owners intervene on uncertainty. Through the Agent SDK, custom taxonomy-driven AI workflows can be built.