1 Comment
User's avatar
User's avatar
Comment removed
Jan 16
Comment removed
Harry Clarkson-Bennett's avatar

Thank you very much! Yeah the more you know about your expected audience, the better your knowledge of the cohort and their lifestyles the better isn't it.

If you can make it work at scale, then connecting to wikidata or the knowledge graphs and disambiguating based on the surrounding content works really well for me.

Just needs some aggressive caching I think for real scake