msokolov commented on issue #13611: URL: https://github.com/apache/lucene/issues/13611#issuecomment-3336667277
It might be good to keep in mind that what's provided by KnnVectorQuery is *approximate* nearest neighbors, and iit is expected that the "true" nearest neighbors will not all be returned all the time. The recall (percent of time you do get the true nearest neighbors) is influenced by a number of factors. One thing you didn't mention is what kind of M or maxconn you used when indexing. Generally speaking increasing that should yield improved recall. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
