The Lemur Project / Feature Requests / #144 Ranklib to save models in Json

Search engine and data mining applications and ClueWeb datasets.

#144 Ranklib to save models in Json

Milestone: Next_Release

Status: open

Owner: Alessandro

Labels: LambdaMART (1) model formats (1)

Priority: 1

Updated: 2016-08-26

Created: 2016-08-24

Creator: Alessandro

Private: No

Apache Solr is almost ready to support Learning To Rank.
In particular, a Bloomberg plugin to support it is in quite an advance status of completion [1] .
The Solr plugin is expecting the model encoded in Json.
Would be really useful to directly produce the model in Json from Ranklib, taking the writer type ( xml or Json) as a parameter.
I can contribute the Json support for LambdaMART.
Any additional contribution is welcome.

[1] Solr LTR Bloomberg plugin - https://issues.apache.org/jira/browse/SOLR-8542

Discussion

Lemur Project - 2016-08-26

labels: LambdaMART --> LambdaMART, model formats

Group: Someday --> Next_Release
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Lemur Project - 2016-08-26

Sounds like a useful feature.

Will try to add for next release.

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Ranklib to save models in Json

Search engine and data mining applications and ClueWeb datasets.

Group

Searches

Help

#144 Ranklib to save models in Json

Discussion