Bringing taxonomic structure to large digital libraries

Digital libraries are invaluable repositories of information. However, in many situations, their size makes it difficult to access the desired resource. In this paper, we present an automatic, unsupervised, domain-independent and scalable approach for structuring the resources available in a certain electronic repository for a particular domain. The system automatically detects and extracts the main topics related to the desired domain, offering a taxonomical structure. This result is complemented by the library’s search engine, offering an integrated tool for accessing resources as an automatically composed directory service. The system has been tested for several digital libraries and domains of knowledge, providing good quality results in all cases.