You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
ideally some amount of fuzzy matching to terms would be automatically used to improve recall. Example would include pluralizations (stromatolite, stromatolites) hyphenations/slashes (stromatolitic-thrombolitic strom/throm).
Fancy option might include flag for explicitly not doing this.
The text was updated successfully, but these errors were encountered:
Could you explain the "optional flag variations"? I'm not sure I follow that one.
Some (hyphens/slashes) we already have in place -- a paper with "stromatolitic-thrombolitic" will match for a term search of "stromatolite". Some of this (pluralizations) we'll get when we next build a new index (which I'm planning to do alongside an upgrade to ES 2.3 when the new servers arrive). Other pieces seem similar in concept to our proposed hierarchy crawler to clean up the signal (obviously requiring other terms in the hierarchy is just a special case of an AND combination).
here is a pretty complex example:
Include documents that match: [term='permeability' OR term='hydraulic conductivity' OR term='transmissivity'] AND [term='St Peter Sandstone' OR term='Tuscaloosa' OR term='Carbondale' OR term='Niobrara' OR term='Mount Simon Sandstone' OR term='Chattanooga Shale’]
ideally some amount of fuzzy matching to terms would be automatically used to improve recall. Example would include pluralizations (stromatolite, stromatolites) hyphenations/slashes (stromatolitic-thrombolitic strom/throm).
Fancy option might include flag for explicitly not doing this.
The text was updated successfully, but these errors were encountered: