Webinar: Shaping the semantic layer by mining digitised data: an encounter between iDigBio's plant records and the Environment Ontology (ENVO)

Add this event to your calendar:

Add to My Calendar
Wed, 2015-10-28 12:54 -- dpaul

UPDATE: Did you miss this? Would you like to hear bits again? The recording captured the community conversation - and ideas for next steps. If you have ideas for furthering this work, please send a note to idigbio@acis.ufl.edu.

What's an ontology? Why would a structured, hierachical group of enviromental terms be useful in enviromental research? How does text-mining habitat data from aggregators and integrators like iDigBio make it possible to create richer ontologies benefitting research?

Abstract: Domain ontologies seek to represent the entities in an area of investigation alongside their interrelations in a logically coherent manner. Domain knowledge is the lifeblood of these ontologies and initiatives such as iDigBio offer valuable repositories which contain both legacy and contemporary knowledge and are ripe for mining. The Environment Ontology (ENVO) is a domain ontology that is concerned with environments, ecosystems, habitats, and related entities. We performed a cursory text-mining exercise focusing on iDigBio's plant habitat descriptions to learn what terms and phrases the botany community - past and present - have used to describe the environments they sample. The results are being used to guide the creation of new classes and relations in ENVO, ensuring the ontology is fit to represent the environmental aspects of these early waves of digitised plant data. In this webinar, I will describe our approach, some of its early results, and the main challenges we face going forward.

There is considerable scope to refine and extend this approach to other digitised data and power synergy between the semantic layer and data repositories.

Presenters:

Dr. Pier Luigi Buttigieg
HGF-MPG Group for Deep Sea Ecology and Technology
Alfred-Wegener-Institut, Helmholtz-Zentrum für Polar- und Meeresforschung
c/o Max Planck Institute for Marine Microbiology
Celsiusstrasse 1, 28359 Bremen, Germany
Email: pbuttigi@mpi-bremen.de

and

Grant Godden
Research Associate, Mint Genome Project
Buell Lab http://buell-lab.plantbiology.msu.edu/index.shtml
Michigan State University
Email: goddengr@msu.edu
Former iDigBio Graduate Student, Soltis Lab

Brought to you by the iDigBio Data Management Interest Group in collaboration with the Cyberinfrastructure Working Group and hosts: Deb Paul, Matt Collins, and Alex Thompson.

When: Tue 9:00 AM EST (Tallahassee, Florida) Tue 3:00 PM CET (Berlin, Germany)

Start Date: 
Tuesday, November 10, 2015 - 9:00am to 10:00am EST
Location: 
Bremen, Germany and East Lansing, MI, USA
AdobeConnect URL: 
http://idigbio.adobeconnect.com/datamgmt