Collaborators
Matt King
Data Science Team
While general purpose search engines can assist with finding research documents relevant to themes of interest, this approach can lack the perspective of experts. RangeDocs is an online tool designed to search a catalog of rangeland science at the paragraph level. Rangeland experts have read the literature and tagged common rangeland terms where they appear in the documents.
We trained a natural language model that using text labeled by rangeland research experts to predicts terms associated with new texts. This allows more efficient tagging. Continuing work is aimed at validating and incorporating papers tagged by these ML models into the RangeDocs search engine.