3

I am working on concept detection and matching for museums and other cultural institutions. Starting with the preloaded Wikipedia graph in Watson "concept insights" is a good starting point. However, when I have large content repositories I want to search and match, I need a graph for these specific corpora. How can this be achieved?

kay.hartkopf
  • 31
  • 4
  • not sure if I've got your question, but the API suggests that you can provide watson your own corpus https://watson-api-explorer.mybluemix.net/apis/concept-insights-v2?cm_mc_uid=75357136835914056170021&cm_mc_sid_50200000= – Leo Feb 26 '16 at 12:51
  • Leo, let me explain a little more. Sure, you can add your own corpus but you do not have access to a graph build for this corpus like the one for Wikipedia. So you can not navigate your own corpus in the same way. This is my understanding so far. The question is, is there a similar graph available for my own corpora and can it be used in the same way the Wikipedia graph can be used for the Wkipedia corpus? – kay.hartkopf Feb 28 '16 at 08:11
  • I see. the answer is: I am afraid not – Leo Feb 29 '16 at 14:34

1 Answers1

0

I am a developer of the Concept Insights service. Thank you for letting us know that the system based on a wikipedia graph provides some degree of value to your use case. We would be very interested in learning what you think more specifically the service should be able to do to better serve in your application. Is it the case that there are concepts that you would like the service to pick up that it currently does not?

Luis A Lastras
  • 44
  • 3
  • If I may add my own answer to this question: I would really like to use some simple concepts that Concept Insights currently does not have. For example: mystery, timeless, cosy, introvert, extravert, hurry, etc. Currently I try to match these concepts to Wikipedia concepts, but that is often not really possible. – gmjonker May 20 '16 at 08:00