An intersection ontology may comprise a root node, representing an entire data set, and a second "layer" of nodes, groups of which may be used to represent different types of classifications of the data set. The intersection ontology may then contain one or more layers of nodes that may be represented by intersections of two or more of the second-layer nodes, as may be needed according to a desired application (that is, the intersections may be formed, essentially, on demand). There may be a single layer of intersection nodes, or there may be multiple layers of intersection nodes. In a multiple layer embodiment, each intersection node may be represented as an intersection of second-layer nodes and/or intersection nodes.