Percolator: Best Practice for Boost/Customizing Relevance Scoring

Per documentation, index-time boosting was replaced with query-time boosting in a previous release. This makes sense - here an explanation is given which explains that index time boost requires re-indexing, so it's better to simply do it at query time, which doesn't require re-indexing to change boost and/or scoring criteria.

I'm now using Percolator and want to boost one property match higher than the others. Based on "index time boosting" being bad (and agreeing with the reasoning), and that percolator queries themselves are indexed, is there any way to boost a property when querying the percolator field with a document that contains that property? Or does the boost need to be saved with the percolator query in this case?

Since the request to percolate only includes property names and their values, I'm not aware of a way to change boost or anything else that would impact relevance scoring without saving it with the query or on the index definition, the latter definitely being deprecated, and the former method seeming to have the same issue with it.

Thank you!

Note that I have also posted this question in the following places:

This may be a relevant post: Understanding Similarity Scoring in Elasticsearch

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.