WG3 - Ontologies and NLP

This working group aims to:

The aim of working group 3 is to build a shared approach towards natural language processing and ontology management. A shared natural language processing effort will increase development efficiency and build consensus around the handling of free-text data in observational cancer-care research. Shared ontological models will ensure that the outputs of this NLP work are well-described and replicable, as well as increasing the reusability of data both in structured and free-text data fields.

The focus of WG3 to date has been to:

• Review technological solutions and select a platform for future development
• Consultation with clinical stakeholders for feasibility and prioritisation of NLP targets
• Proof of concept pipeline delivery and implementation of shared processing library and architecture, as well as basic named entity recognition functionality
• Knowledge-sharing around conceptual and technological aspects of work done to date

Next steps:

• Implement highest priority clinical targets delivered by clinical consultation step
• Review clinical targets to select sensible scope of the first phase of ontological harmonisation
• Build out knowledge book entries on top of NLP pipeline to use selected ontologies
• Define sustainable ongoing processes for handling future development and mapping efforts