Conservation Evidence Group
Phil Martin <pam79@cam.ac.uk>
Curating Conservation Evidence
A core activity of the Cambridge Conservation Science Group is compiling evidence from research literature, to show which management approaches are most effective for biodiversity conservation. However, this evidence is published in many different academic fields. The Group need tools that can use natural language processing methods to constantly monitor publications across many different venues, extracting key attributes such as geography, habitat, threats or interventions and organising these for thematic browsing or targeted queries without being constrained by the peculiar formats or terminology of particular scientific disciplines.
Original query:
I'm part of the Conservation Evidence team in Zoology, where we work on summarising scientific evidence on the effectiveness of different types of management for biodiversity conservation.
As part of this process we systematically trawl through the scientific literature to identify publications that are relevant to us, which can be extremely time-consuming. We're interested in seeing if someone could use a machine-learning approach to identify relevant studies based on their contents. We have a database that contains studies that we have already identified as relevant that could be used to train the algorithm. Ideally the tool that would be produced would also classify the topic of the studies. We estimate that a tool such as this could save us ~5-15% of the time taken to synthesise evidence.
Does this sound like a good idea for a group project?
Regards,
-- Phil Martin
Postdoctoral Research Associate
Conservation Evidence
Conservation Science Group – University of Cambridge
