Many of the institutions providing metadata to the DPLA organize their holdings in collections. Some collection details supplied by the data provider, such as the title and description, are retained in the item records but not displayed. The Computational Collection Descriptions project, hosted by the School of Information at the University of Texas at Austin, is developing mechanisms that use item-level data to generate collection-level attributes such as subjects, dates, and locations. Enhanced collection-level metadata will enable the discovery of whole collections, support relevance assessment of items in the context of collections, help manage search results by grouping at the collection level, and provide greater visibility for results from institutions that have fewer matching objects. Creating collection-level views of the aggregated metadata will also support new approaches to the evaluation and enrichment of metadata at large scale.
Notes available at
https://drive.google.com/open?id=1XKQdU1RHXU69A6BwDWMwLCroH6th9HpnItTrIuNM0lY&authuser=1
