[SPARK-1260]: faster construction of features with intercept

The current implementation uses `Array(1.0, features: _*)` to construct a new array with intercept. This is not efficient for big arrays because `Array.apply` uses a for loop that iterates over the arguments. `Array.+:` is a better choice here.

Also, I don't see a reason to set initial weights to ones. So I set them to zeros.

JIRA: https://spark-project.atlassian.net/browse/SPARK-1260

Author: Xiangrui Meng <meng@databricks.com>

Closes #161 from mengxr/sgd and squashes the following commits:

b5cfc53 [Xiangrui Meng] set default weights to zeros
a1439c2 [Xiangrui Meng] faster construction of features with intercept
This commit is contained in:
Xiangrui Meng 2014-03-18 15:14:13 -07:00 committed by Reynold Xin
parent 79e547fe5a
commit e108b9ab94

View file

@ -119,7 +119,7 @@ abstract class GeneralizedLinearAlgorithm[M <: GeneralizedLinearModel]
*/
def run(input: RDD[LabeledPoint]) : M = {
val nfeatures: Int = input.first().features.length
val initialWeights = Array.fill(nfeatures)(1.0)
val initialWeights = new Array[Double](nfeatures)
run(input, initialWeights)
}
@ -134,15 +134,15 @@ abstract class GeneralizedLinearAlgorithm[M <: GeneralizedLinearModel]
throw new SparkException("Input validation failed.")
}
// Add a extra variable consisting of all 1.0's for the intercept.
// Prepend an extra variable consisting of all 1.0's for the intercept.
val data = if (addIntercept) {
input.map(labeledPoint => (labeledPoint.label, Array(1.0, labeledPoint.features:_*)))
input.map(labeledPoint => (labeledPoint.label, labeledPoint.features.+:(1.0)))
} else {
input.map(labeledPoint => (labeledPoint.label, labeledPoint.features))
}
val initialWeightsWithIntercept = if (addIntercept) {
Array(1.0, initialWeights:_*)
initialWeights.+:(1.0)
} else {
initialWeights
}