Lucene Score results

In Lucene if you had numerous indexes that covered just one dividing each. Why does the very same search on various indexes return outcomes with various ratings? The arise from various web servers match specifically.

i.e. if I looked for :

  • Name - John Smith
  • DOB - 11/11/1934

Dividing 0 would certainly return a rating of 0.345

Dividing 1 would certainly return a rating of 0.337

Both suit specifically on name and also DOB.

2019-05-03 23:36:11
Source Share
Answers: 3

You might additionally want the result of the explain(), and also the resulting Explanation, which will certainly offer you a suggestion of just how points are racked up the means they are.

2019-05-09 08:14:35

The scoring has the Inverse Document Frequency (IDF ). If the term "John Smith" remains in one dividing, 0, 100 times and also in dividing 1, as soon as. Ball game for looking for John Smith would certainly be greater search in dividing 1 as the term is extra limited.

To get round this you would certainly perish need to have your index moring than all dividings, or you would certainly require to bypass the IDF.

2019-05-07 17:19:59

Because ball game is established on the index if I am not entirely misinterpreted.

If you have various indexes (more/less or various information that was indexed ), ball game will certainly vary :

(Warning : Contains Math :- ) )

2019-05-07 16:52:37