Font Size:
Prototyping a botanical knowledge portal
Building: Grand Hotel Mediterraneo
Room: Sala dei Continenti
Date: 2013-11-01 11:52 AM – 12:01 PM
Last modified: 2013-10-07
Abstract
Canada has not published a national Flora since 1979 and, since this time, there have been considerable changes in diversity both from loss due to habitat destruction, and increases from invasive and alien introductions. This illustrates a general problem in Flora publication, which is that they are often out of date long before they are officially revised. We are building a botanical knowledge portal that seeks to address this problem, and, at the same time, provide character-based querying over treatment content.
We are doing this by using the CharaParser toolset to do fine-grained semantic parsing of Flora of North America treatments, and to express the resulting structured knowledge as Extensible Markup Language: XML. We produced a demonstration interface to this data by converting the XML to Resource Description Framework or RDF for loading into Jena, and are now building on this experience by using the semantic extensions to MediaWiki (Semantic MediaWiki and WikiData) for storing structured content. Floras are dynamic by nature and this platform will enable both updating and correction of mistakes, as well as integration of information from other sources, including specimens and observations, annotations, checklists, images and illustrations, maps, genetic resources, and other literature sources.
To enable this, our recent focus has been to produce extension ontologies focusing on FNA terms, in order to broker semantic comparison across treatments, and, through links to other ontologies, foster integration with other resources. When linked to relevant glossaries/ontologies these term building blocks and their attributes can be used to create revised/updated descriptions, as markup for regions-of-interest on images, components of an interactive key created on-the-fly based on characteristics relevant to identifying species, etc.
We will highlight our progress on this botanical knowledge portal by briefly describing its components and workflow, which includes parsing the treatments, crowdsourcing the validation and correction of the output, and loading structured data into a platform that enables semantic querying and integration. We will demonstrate prototype portal capabilities, and seek feedback to inform our ongoing work.
We are doing this by using the CharaParser toolset to do fine-grained semantic parsing of Flora of North America treatments, and to express the resulting structured knowledge as Extensible Markup Language: XML. We produced a demonstration interface to this data by converting the XML to Resource Description Framework or RDF for loading into Jena, and are now building on this experience by using the semantic extensions to MediaWiki (Semantic MediaWiki and WikiData) for storing structured content. Floras are dynamic by nature and this platform will enable both updating and correction of mistakes, as well as integration of information from other sources, including specimens and observations, annotations, checklists, images and illustrations, maps, genetic resources, and other literature sources.
To enable this, our recent focus has been to produce extension ontologies focusing on FNA terms, in order to broker semantic comparison across treatments, and, through links to other ontologies, foster integration with other resources. When linked to relevant glossaries/ontologies these term building blocks and their attributes can be used to create revised/updated descriptions, as markup for regions-of-interest on images, components of an interactive key created on-the-fly based on characteristics relevant to identifying species, etc.
We will highlight our progress on this botanical knowledge portal by briefly describing its components and workflow, which includes parsing the treatments, crowdsourcing the validation and correction of the output, and loading structured data into a platform that enables semantic querying and integration. We will demonstrate prototype portal capabilities, and seek feedback to inform our ongoing work.