Search code examples
watson-discoverywatson-knowledge-studio

Importing previously annotated document to IBM Knowledge Studio


I am doing some research on building custom models for Entity extraction. For this i have used some of the public dataset and wanted to see how it performs in IBM Knowledge Studio. But i am finding it difficult to find a way to load the public dataset (which is already annotated) to the Knowledge Studio.

There document says, previously annotated documents can be imported, but it doesn't specify about the format https://console.bluemix.net/docs/services/watson-knowledge-studio/create-project.html#create-project

Document also says, it can be from a UIMA analysis engine, i don't find any good examples which show the format of the file

Can anyone help on this?


Solution

  • Hope this helps other.

    There is no clear documentation on the input format when you are trying to import existing annotated data in to Knowledge Studio.

    We did a work around, by manually annotating few files from Knowledge Studio and exported this data. We did analysis on this exported data and wrote custom programs which will transform the existing annotated text to the format that Knowledge Studio accepts.

    Then imported this data back to Knowledge Studio