diff --git a/mllib/src/main/scala/org/apache/spark/mllib/evaluation/binary/BinaryClassificationMetrics.scala b/mllib/src/main/scala/org/apache/spark/mllib/evaluation/binary/BinaryClassificationMetrics.scala
index c95cbf525a2e8..658142b0ebf53 100644
--- a/mllib/src/main/scala/org/apache/spark/mllib/evaluation/binary/BinaryClassificationMetrics.scala
+++ b/mllib/src/main/scala/org/apache/spark/mllib/evaluation/binary/BinaryClassificationMetrics.scala
@@ -17,7 +17,7 @@
 
 package org.apache.spark.mllib.evaluation.binary
 
-import org.apache.spark.rdd.RDD
+import org.apache.spark.rdd.{UnionRDD, RDD}
 import org.apache.spark.SparkContext._
 import org.apache.spark.mllib.evaluation.AreaUnderCurve
 import org.apache.spark.Logging
@@ -103,10 +103,17 @@ class BinaryClassificationMetrics(scoreAndLabels: RDD[(Double, Double)])
 
   /**
    * Returns the receiver operating characteristic (ROC) curve,
-   * which is an RDD of (false positive rate, true positive rate).
+   * which is an RDD of (false positive rate, true positive rate)
+   * with (0.0, 0.0) prepended and (1.0, 1.0) appended to it.
    * @see http://en.wikipedia.org/wiki/Receiver_operating_characteristic
    */
-  def roc(): RDD[(Double, Double)] = createCurve(FalsePositiveRate, Recall)
+  def roc(): RDD[(Double, Double)] = {
+    val rocCurve = createCurve(FalsePositiveRate, Recall)
+    val sc = confusions.context
+    val first = sc.makeRDD(Seq((0.0, 0.0)), 1)
+    val last = sc.makeRDD(Seq((1.0, 1.0)), 1)
+    new UnionRDD[(Double, Double)](sc, Seq(first, rocCurve, last))
+  }
 
   /**
    * Computes the area under the receiver operating characteristic (ROC) curve.
@@ -114,11 +121,16 @@ class BinaryClassificationMetrics(scoreAndLabels: RDD[(Double, Double)])
   def areaUnderROC(): Double = AreaUnderCurve.of(roc())
 
   /**
-   * Returns the precision-recall curve,
-   * which is an RDD of (recall, precision), NOT (precision, recall).
+   * Returns the precision-recall curve, which is an RDD of (recall, precision),
+   * NOT (precision, recall), with (0.0, 1.0) prepended to it.
    * @see http://en.wikipedia.org/wiki/Precision_and_recall
    */
-  def pr(): RDD[(Double, Double)] = createCurve(Recall, Precision)
+  def pr(): RDD[(Double, Double)] = {
+    val prCurve = createCurve(Recall, Precision)
+    val sc = confusions.context
+    val first = sc.makeRDD(Seq((0.0, 1.0)), 1)
+    first.union(prCurve)
+  }
 
   /**
    * Computes the area under the precision-recall curve.
diff --git a/mllib/src/test/scala/org/apache/spark/mllib/evaluation/binary/BinaryClassificationMetricsSuite.scala b/mllib/src/test/scala/org/apache/spark/mllib/evaluation/binary/BinaryClassificationMetricsSuite.scala
index a1f3c44becb66..f92adbe8378c6 100644
--- a/mllib/src/test/scala/org/apache/spark/mllib/evaluation/binary/BinaryClassificationMetricsSuite.scala
+++ b/mllib/src/test/scala/org/apache/spark/mllib/evaluation/binary/BinaryClassificationMetricsSuite.scala
@@ -35,15 +35,16 @@ class BinaryClassificationMetricsSuite extends FunSuite with LocalSparkContext {
     val precision = tp.zip(fp).map { case (t, f) => t.toDouble / (t + f) }
     val recall = tp.map(t => t.toDouble / p)
     val fpr = fp.map(f => f.toDouble / n)
-    val roc = fpr.zip(recall)
+    val rocCurve = Seq((0.0, 0.0)) ++ fpr.zip(recall) ++ Seq((1.0, 1.0))
     val pr = recall.zip(precision)
+    val prCurve = Seq((0.0, 1.0)) ++ pr
     val f1 = pr.map { case (re, prec) => 2.0 * (prec * re) / (prec + re) }
     val f2 = pr.map { case (re, prec) => 5.0 * (prec * re) / (4.0 * prec + re)}
     assert(metrics.thresholds().collect().toSeq === score)
-    assert(metrics.roc().collect().toSeq === roc)
-    assert(metrics.areaUnderROC() === AreaUnderCurve.of(roc))
-    assert(metrics.pr().collect().toSeq === pr)
-    assert(metrics.areaUnderPR() === AreaUnderCurve.of(pr))
+    assert(metrics.roc().collect().toSeq === rocCurve)
+    assert(metrics.areaUnderROC() === AreaUnderCurve.of(rocCurve))
+    assert(metrics.pr().collect().toSeq === prCurve)
+    assert(metrics.areaUnderPR() === AreaUnderCurve.of(prCurve))
     assert(metrics.fMeasureByThreshold().collect().toSeq === score.zip(f1))
     assert(metrics.fMeasureByThreshold(2.0).collect().toSeq === score.zip(f2))
     assert(metrics.precisionByThreshold().collect().toSeq === score.zip(precision))