Database evaluation with data dumps
If you use Freebase search to find matches for any item by name, you will get results sorted by relevance score . Try the Taj Mahal , for example .
I'm trying to get similar results using Freebase dumps data, so in my database, Taj Mahal related questions will be sorted by relevance, meaning building comes first, musician comes next, and so on.
Are there any suggestions for achieving this without requesting the Freebase search API?
source to share
The relevancy wiki page you linked to says:
Freebase objects have a built-in relevance score (ranking) calculated during indexing, which is a function of the number of incoming and outgoing links in Freebase and Wikipedia. Some popular Freebase organizations also have Google Popularity Rankings. By default, both accounts are pooled at the time of queries.
Which should give you a pretty good idea of where to start. Freebase in-degree and out-degree can be computed directly from the dump, but Wikipedia in / out of bounds requires a Wikipedia dump (or Freebase WEX dump). Google's "Popularity Score" is obviously something you can't replicate.
source to share