[
https://issues.apache.org/jira/browse/SPARK-18205?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15626396#comment-15626396
]
Sean Owen commented on SPARK-18205:
-----------------------------------
Could you give a more specific example of what you have in mind? "more docs for
all transformers" is too broad. Models produce some additional output given
input and so all transform methods are generally adding columns to a DataFrame
containing some thing they predict or estimate.
> Spark ML Estimators and Models needs documentation
> ---------------------------------------------------
>
> Key: SPARK-18205
> URL: https://issues.apache.org/jira/browse/SPARK-18205
> Project: Spark
> Issue Type: Bug
> Components: Documentation
> Affects Versions: 2.0.3
> Reporter: nirav patel
>
> I see that most of the Algorithms are missing documentation. It's really hard
> to know what exactly Fit and Transform methods does for given algorithm
> without it. e.g. One can assume that IDF model Fit method will compute TF-IDF
> for given dataset. but then how it uses that information to Transform new
> dataset? In other words, what parameters exactly Estimator learned during
> Fitting and will re-use it for Transformation.
> Same thing apply for all Estimators and Transformers.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
