sul-dlss-labs / spoc

Species Occurrences (SpOc), documentation available at https://sul-dlss-labs.github.io/spoc/
2 stars 0 forks source link

Entity matching not using full entity? #73

Open amandawhitmire opened 3 years ago

amandawhitmire commented 3 years ago

In some cases, we are seeing entity "hits" on words that are likely only part of an entity. For example, "beach" is tagged as a location in the text, when the entity set had named location matches that are "Hopkins Beach" or "Moss Beach". "Beach," by itself, should not be tagged as an entity in the NER process.