A technique is described for performing domain-specific analysis, structuring, mapping and classification of data entities, such as text document, images, audio data, waveform data, and so forth. A domain definition is established that includes a plurality of classification axes and labels for each axis. Data entities are accessed that potentially have attributes of interest classifiable in accordance with the axes and labels. Pertinent entities are then identified based upon their attributes, and the entities are classified. The classification and the entities themselves, or portions thereof, may be stored in a knowledge base for further classification, search and reference. Complex combinations of classifications, including combinations by reference to data of different type are possible by virtue of the domain definition and rules or algorithms called on by the definition for one-to-many mapping of the entities to the axes and labels.