Applying Text Mining Services to Facilitate 
Discovery and Linking of Wheat Scientific 
Information and Data
@openminted_eu
Agroknow
Applying Text Mining Services to 
Facilitate Discovery and Linking of 
Wheat Scientific Information and 
Data
The scope of this presentation is to demonstrate 
the process of applying text mining services to 
support discovery and inter-linking of wheat 
scientific information. This is supported by a set 
of useful endpoints that can be used for in this 
process
2
OPENMINTED - The Open Mining Infrastructure for Text and Data
Specific Use Cases
• Let us consider a real-world problem.
• Consider an organization (perhaps yours?) having 
research information and data in different databases.
• How could we connect these data silos?
@openminted_eu
• Need to:
 Define a workflow,
 Design a data model,
 Implement it!
OPENMINTED - The Open Mining Infrastructure for Text and Data
Proposed Solution
• Develop a layer, running on top of the data silos:
 Harvesting the data stored,
 Aligning them in a uniform internal format,
 Enriching/Interlinking them with external systems,
@openminted_eu
 Indexing the enriched data,
 Providing them back through a search api.
OPENMINTED - The Open Mining Infrastructure for Text and Data
Proposed Solution – Complete Picture
Need a data 
model to 
support our 
@openminted_eu
workflow
OPENMINTED - The Open Mining Infrastructure for Text and Data
Data Model (1/2)
• Everything is an object.
• Each object has a specific type, with 
different properties.
@openminted_eu
• Of course everything is interlinked.
• Any new content type can be added on 
the second level with specific  properties.
OPENMINTED - The Open Mining Infrastructure for Text and Data
Data Model (2/2)
@openminted_eu
OPENMINTED - The Open Mining Infrastructure for Text and Data
Enrichment Process
• The process: 
 use raw information stored in resources (title, 
abstract, author/publisher list, full-text etc.), 
 recognize entities using various endpoints, 
@openminted_eu
 interlinking them with both internal entities 
and external systems.
OPENMINTED - The Open Mining Infrastructure for Text and Data
Enrichment Process – Initial State
@openminted_eu
OPENMINTED - The Open Mining Infrastructure for Text and Data
Enrichment Process – Entity Recognition
@openminted_eu
OPENMINTED - The Open Mining Infrastructure for Text and Data
Enrichment Process – Entity Interlinking
@openminted_eu
Some statistics using this process (and only 3 endpoints):
Useful Endpoints (1/2)
• FREME API, used for topics extraction and 
annotation (against AGROVOC), and entity 
recognition (person, organization, location),
• Geonames API, for location extraction and 
interlink,
• OpenAIRE mining service (part of the 
OpenMinTed project), can be used to mine 
projects from text, data citation, classification, 
etc.
@openminted_eu
Useful Endpoints (2/2)
• CropOntology, can be used to extract wheat 
trait entities,
• PDF Text Extraction (and annotation) service, 
used to extract text from pdf files and annotate 
it using various endpoints (1st prize in 1st 
AgroHackathon, Montpellier, 29/6-1/7/2016)
@openminted_eu
Outcomes Using these Technologies
• AKIF Search API
• CIMMYT MetaSearch API
@openminted_eu
Analytics (1/3)
@openminted_eu
Analytics (2/3)
@openminted_eu
Analytics (3/3)
@openminted_eu
twitter.com/openminted_eu
facebook.com/openminted
bit.do/openmintedlinkedin
vimeo.com/openminted
bit.do/openmintedplus
www.openminted.eu