Removed dependency on Apache Commons Math

5 files changed, 27 insertions, 27 deletions

diff --git a/ml/src/main/scala/spark/ml/regression/LogisticRegressionGenerator.scala b/ml/src/main/scala/spark/ml/regression/LogisticRegressionGenerator.scala
index 1617eac205..6d37aad047 100644
--- a/ml/src/main/scala/spark/ml/regression/LogisticRegressionGenerator.scala
+++ b/ml/src/main/scala/spark/ml/regression/LogisticRegressionGenerator.scala
@@ -1,11 +1,12 @@
 package spark.ml.regression
 
-import spark.{RDD, SparkContext}
-import spark.ml.util.MLUtils
+import scala.util.Random
 
-import org.apache.commons.math3.distribution.NormalDistribution
 import org.jblas.DoubleMatrix
 
+import spark.{RDD, SparkContext}
+import spark.ml.util.MLUtils
+
 object LogisticRegressionGenerator {
 
   def main(args: Array[String]) {
@@ -25,12 +26,11 @@ object LogisticRegressionGenerator {
     val sc = new SparkContext(sparkMaster, "LogisticRegressionGenerator")
 
     val data: RDD[(Double, Array[Double])] = sc.parallelize(0 until nexamples, parts).map { idx =>
-      val rnd = new NormalDistribution(0, 1)
-      rnd.reseedRandomGenerator(42 + idx)
+      val rnd = new Random(42 + idx)
 
       val y = if (idx % 2 == 0) 0 else 1
       val x = Array.fill[Double](nfeatures) {
-        rnd.sample() + (y * eps)
+        rnd.nextGaussian() + (y * eps)
       }
       (y, x)
     }
diff --git a/ml/src/main/scala/spark/ml/regression/RidgeRegressionGenerator.scala b/ml/src/main/scala/spark/ml/regression/RidgeRegressionGenerator.scala
index ac7f1e7320..75854fe1de 100644
--- a/ml/src/main/scala/spark/ml/regression/RidgeRegressionGenerator.scala
+++ b/ml/src/main/scala/spark/ml/regression/RidgeRegressionGenerator.scala
@@ -1,11 +1,12 @@
 package spark.ml.regression
 
-import spark.{RDD, SparkContext}
-import spark.ml.util.MLUtils
+import scala.util.Random
 
-import org.apache.commons.math3.distribution.NormalDistribution
 import org.jblas.DoubleMatrix
 
+import spark.{RDD, SparkContext}
+import spark.ml.util.MLUtils
+
 
 object RidgeRegressionGenerator {
 
@@ -38,10 +39,9 @@ object RidgeRegressionGenerator {
       val X = DoubleMatrix.rand(examplesInPartition, nfeatures)
       val y = X.mmul(w)
 
-      val rnd = new NormalDistribution(0, eps)
-      rnd.reseedRandomGenerator(42 + p)
+      val rnd = new Random(42 + p)
 
-      val normalValues = Array.fill[Double](examplesInPartition)(rnd.sample())
+      val normalValues = Array.fill[Double](examplesInPartition)(rnd.nextGaussian() * eps)
       val yObs = new DoubleMatrix(normalValues).addi(y)
 
       Iterator.tabulate(examplesInPartition) { i =>
diff --git a/ml/src/test/scala/spark/ml/regression/LogisticRegressionSuite.scala b/ml/src/test/scala/spark/ml/regression/LogisticRegressionSuite.scala
index ce388ecc26..53d9789979 100644
--- a/ml/src/test/scala/spark/ml/regression/LogisticRegressionSuite.scala
+++ b/ml/src/test/scala/spark/ml/regression/LogisticRegressionSuite.scala
@@ -1,22 +1,23 @@
 package spark.ml.regression
 
+import scala.util.Random
+
+import org.scalatest.FunSuite
+
 import spark.SparkContext
 import spark.SparkContext._
 import spark.Logging
 
-import org.apache.commons.math3.distribution.NormalDistribution
-import org.scalatest.FunSuite
 
 class LogisticRegressionSuite extends FunSuite with Logging {
 
   // Test if we can correctly learn A, B where Y = logistic(A + B*X)
   test("logistic regression") {
     val nPoints = 10000
-    val rnd = new NormalDistribution(0, 1)
-    rnd.reseedRandomGenerator(42)
+    val rnd = new Random(42)
 
     val sc = new SparkContext("local", "test")
-    val x1 = Array.fill[Double](nPoints)(rnd.sample())
+    val x1 = Array.fill[Double](nPoints)(rnd.nextGaussian())
 
     val A = 2.0
     val B = -1.5
diff --git a/ml/src/test/scala/spark/ml/regression/RidgeRegressionSuite.scala b/ml/src/test/scala/spark/ml/regression/RidgeRegressionSuite.scala
index 6d5f13d6f5..795cda1379 100644
--- a/ml/src/test/scala/spark/ml/regression/RidgeRegressionSuite.scala
+++ b/ml/src/test/scala/spark/ml/regression/RidgeRegressionSuite.scala
@@ -1,25 +1,25 @@
 package spark.ml.regression
 
+import scala.util.Random
+
+import org.scalatest.FunSuite
+
 import spark.SparkContext
 import spark.SparkContext._
 
-import org.apache.commons.math3.distribution.NormalDistribution
-import org.scalatest.FunSuite
 
 class RidgeRegressionSuite extends FunSuite {
 
   // Test if we can correctly learn Y = 3 + X1 + X2 when
   // X1 and X2 are collinear.
   test("multi-collinear variables") {
-    val rnd = new NormalDistribution(0, 1)
-    rnd.reseedRandomGenerator(43)
+    val rnd = new Random(43)
     val sc = new SparkContext("local", "test")
-    val x1 = Array.fill[Double](20)(rnd.sample())  
+    val x1 = Array.fill[Double](20)(rnd.nextGaussian())
 
     // Pick a mean close to mean of x1
-    val rnd1 = new NormalDistribution(0.1, 0.01)
-    rnd1.reseedRandomGenerator(42)
-    val x2 = Array.fill[Double](20)(rnd1.sample())
+    val rnd1 = new Random(42) //new NormalDistribution(0.1, 0.01)
+    val x2 = Array.fill[Double](20)(0.1 + rnd1.nextGaussian() * 0.01)
 
     val xMat = (0 until 20).map(i => Array(x1(i), x2(i))).toArray
 
diff --git a/project/SparkBuild.scala b/project/SparkBuild.scala
index aa877ad4a7..5dbb5d4a65 100644
--- a/project/SparkBuild.scala
+++ b/project/SparkBuild.scala
@@ -224,8 +224,7 @@ object SparkBuild extends Build {
   def mlSettings = sharedSettings ++ Seq(
     name := "spark-ml",
     libraryDependencies ++= Seq(
-      "org.jblas" % "jblas" % "1.2.3",
-      "org.apache.commons" % "commons-math3" % "3.2"
+      "org.jblas" % "jblas" % "1.2.3"
     )
   )