Open stripathy opened 8 years ago
Plan - take the 1K neuroelectro hand curated spans of neuron mentions and run through various ontologies - in how many cases were the instances perfectly matched?
A more unbiased way to do this would be to use curated neuron names from some other resource, like neuromorpho
The advantages of neuroNER are that it allows finer granularity for describing neuron instances than currently provided by neuroLEX, Cell Ontology, etc (from my experience curating against the literature for NeuroElectro).
We need to make a set of figures clearly illustrating this.