[SPARK-6496] [MLLIB] GeneralizedLinearAlgorithm.run(input, initialWeights) should initialize numFeatures

In GeneralizedLinearAlgorithm ```numFeatures``` is default to -1, we need to update it to correct value when we call run() to train a model.
```LogisticRegressionWithLBFGS.run(input)``` works well, but when we call ```LogisticRegressionWithLBFGS.run(input, initialWeights)``` to train multiclass classification model, it will throw exception due to the numFeatures is not updated.
In this PR, we just update numFeatures at the beginning of GeneralizedLinearAlgorithm.run(input, initialWeights) and add test case.

Author: Yanbo Liang <ybliang8@gmail.com>

Closes #5167 from yanboliang/spark-6496 and squashes the following commits:

8131c48 [Yanbo Liang] LogisticRegressionWithLBFGS.run(input, initialWeights) should initialize numFeatures

1 files changed, 6 insertions, 0 deletions

diff --git a/mllib/src/test/scala/org/apache/spark/mllib/classification/LogisticRegressionSuite.scala b/mllib/src/test/scala/org/apache/spark/mllib/classification/LogisticRegressionSuite.scala
index aaa81da9e2..a26c52852c 100644
--- a/mllib/src/test/scala/org/apache/spark/mllib/classification/LogisticRegressionSuite.scala
+++ b/mllib/src/test/scala/org/apache/spark/mllib/classification/LogisticRegressionSuite.scala
@@ -425,6 +425,12 @@ class LogisticRegressionSuite extends FunSuite with MLlibTestSparkContext with M
 
     val model = lr.run(testRDD)
 
+    val numFeatures = testRDD.map(_.features.size).first()
+    val initialWeights = Vectors.dense(new Array[Double]((numFeatures + 1) * 2))
+    val model2 = lr.run(testRDD, initialWeights)
+
+    LogisticRegressionSuite.checkModelsEqual(model, model2)
+
     /**
      * The following is the instruction to reproduce the model using R's glmnet package.
      *