spark-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Kai Sasaki (JIRA)" <>
Subject [jira] [Commented] (SPARK-11439) Optimization of creating sparse feature without dense one
Date Fri, 13 Nov 2015 01:48:11 GMT


Kai Sasaki commented on SPARK-11439:

It seems to indicate the model in SparkR here. According to this documentation, you can create
SparkR linear model with `glm`.

This will call {{SparkRWrapper#fitRModelFormula}. It returns LinearRegressionModel with Pipeline
when it receives "gaussian" as second parameter.
So in summary we can write the code like this to use {{LinearRegressionModel}} in SparkR.
df <- createDataFrame(sqlContext, iris)
fit <- glm(Sepal_Length ~ Sepal_Width + Species, data = df, family = "gaussian")

In my environment, it seems to work.

> Optimization of creating sparse feature without dense one
> ---------------------------------------------------------
>                 Key: SPARK-11439
>                 URL:
>             Project: Spark
>          Issue Type: Improvement
>          Components: ML
>            Reporter: Kai Sasaki
>            Priority: Minor
> Currently, sparse feature generated in {{LinearDataGenerator}} needs to create dense
vectors once. It is cost efficient to prevent from generating dense feature when creating
sparse features.

This message was sent by Atlassian JIRA

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message