Pingar Research: One-Click Taxonomy Extraction
HAMILTON, NZ (28 September 2011): Pingar formally announced today at the University of Waikato that it had begun a new research program designed to automate the extraction of taxonomies from unstructured enterprise data.
A number of leading international Natural Language Processing, Machine Learning and Data Mining researchers are engaged in the program.
The Pingar research team is led by Pingar’s Chief Research Officer, Dr. Alyona Medelyan, an expert in Natural Language Processing, developing algorithms that use taxonomies as a source for metadata in searching. Dr. Jeen Broekstra, Pingar consultant, is an expert in Semantic Web & Linked Data. Steve Manion, enrolled in a PhD at Canterbury University, interned at Pingar, has a strong Machine Learning & Natural Language Processing background.
The University of Waikato project lead is Professor Ian Witten, Head of Digital Libraries Lab and receiver of the 2010 World Class New Zealand Research, Science, Technology and Academia Award. Dr. Dave Milne, currently engaged in Post-Doctoral research is an expert in Wikipedia Mining and search user interfaces. Dr. Anna Huang, also currently engaged in Post-Doctoral research is an expert in document clustering and Machine Learning. The goal of the research is to develop algorithms that extract accurate taxonomies on a par with those produced by skilled digital librarians. It addresses a key pain point for enterprises; the need to develop custom taxonomies. The commercial application that Pingar expects to be the outcome of this research will enable organizations to dynamically generate taxonomies from their own documents. These taxonomies will in turn improve how these documents are organized and retrieved, and make it easier to extract information and identify patterns across them.
During the visit to the University of Waikato in Hamilton, New Zealand, Peter Wren-Hilton, Pingar CEO, accompanied by the Earl of Erroll, Chairman of Pingar’s UK Advisory Board, commented:
”‘Following on from the successful commercial deployment of Pingar’s ‘one-click’ metadata assignment set of components for managing unstructured data documents, it became clear that the challenge of building custom taxonomies for enterprises was one that Pingar’s research team had to address. This joint research collaboration with the University of Waikato recognizes the international leadership role that the University of Waikato has in the fields of data mining, natural language processing & machine learning.”
The research program is receiving significant funding from both Pingar and the New Zealand Ministry of Science & Innovation. (www.msi.govt.nz)
