Answer the question
In order to leave comments, you need to log in
Elasticsearch: how to normalize search results by scoring?
We perform sequential search in several arrays. The original request is the same. But the search arrays are different, incl. by the number of objects in them. As a result, the calculated scoring when compared with the same values, but placed in different arrays, turns out to be different: for example, the query contains the word "SLOGAN", a search in an array of 1 million objects gives the same "SLOGAN" with a scoring of 11, and the search in another array of 10k objects it gives "SLOGAN" with a score of 10. That's not to mention the values, which are different. It is difficult to compare search results in the end.
I would like to normalize the results so that in both cases a certain percentage of similarity is returned. In the above example, I would like to get 100% in both cases, because the results are identical to the query.
Tell me where to dig?
Answer the question
In order to leave comments, you need to log in
Didn't find what you were looking for?
Ask your questionAsk a Question
731 491 924 answers to any question