dnmilne / wikipediaminer

An open source toolkit for mining Wikipedia
130 stars 62 forks source link

Questions regarding the API #26

Open shatu opened 9 years ago

shatu commented 9 years ago

Hi guys,

I am doing some study on named entity annotation systems, and would be really glad if you can answer the following questions :-

  1. Is it possible to get NIL mentions i.e. those named entity mentions (surface forms) which WikipediaMiner couldn't link to its Knowledge-Base (either because of lack of sufficient confidence or because the target entity is out of its KB)? Note that, I'm asking this after setting the threshold (minProbability) to zero.
  2. Is it possible to get a ranked and scored list of annotations per mention (surface form)? Will setting the disambiguationPolicy to "loose" return "all possible" candidates for that mention with their corresponding scores/probabilities?
  3. Does your API support specifying a list of mentions (surface names) to annotate?

Thanks, Shashank