You are viewing a plain text version of this content. The canonical link for it is here.

Posted to oak-issues@jackrabbit.apache.org by "Tommaso Teofili (JIRA)" <ji...@apache.org> on 2018/12/13 16:00:00 UTC

[jira] [Created] (OAK-7961) LSH should truncate to first decimal place

Tommaso Teofili created OAK-7961:
------------------------------------

             Summary: LSH should truncate to first decimal place
                 Key: OAK-7961
                 URL: https://issues.apache.org/jira/browse/OAK-7961
             Project: Jackrabbit Oak
          Issue Type: Task
          Components: lucene
            Reporter: Tommaso Teofili
             Fix For: 1.9.14


Feature vector shingles currently contain three decimal places, e.g. _[0.101, 0.202, 0.303]_.
Since min hash is not aware of the fact that second and third decimal place values are less important than first place values, it is likely that this will impact the search precision.
We should be truncating to the first decimal place.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)