[MLLIB] Add fitIntercept param to logistic regression

Made the trait default true
Changed float comparisons to === in unit tests

Author: Omede Firouz

mllib/src/main/scala/org/apache/spark/ml/classification/LogisticRegression.scala

@@ -46,7 +46,6 @@ class LogisticRegression
   with LogisticRegressionParams {
 
   setRegParam(0.1)
-  setFitIntercept(true)
   setMaxIter(100)
   setThreshold(0.5)
 

mllib/src/main/scala/org/apache/spark/ml/param/sharedParams.scala

@@ -112,7 +112,7 @@ private[ml] trait HasFitIntercept extends Params {
    * @group param
    */
   val fitIntercept: BooleanParam =
-    new BooleanParam(this, "fitIntercept", "indicates whether to fit an intercept term")
+    new BooleanParam(this, "fitIntercept", "indicates whether to fit an intercept term", Some(true))
 
   /** @group getParam */
   def getFitIntercept: Boolean = get(fitIntercept)

mllib/src/test/scala/org/apache/spark/ml/classification/LogisticRegressionSuite.scala

@@ -56,14 +56,14 @@ class LogisticRegressionSuite extends FunSuite with MLlibTestSparkContext {
     assert(model.getPredictionCol == "prediction")
     assert(model.getRawPredictionCol == "rawPrediction")
     assert(model.getProbabilityCol == "probability")
-    assert(model.intercept != 0.0)
+    assert(model.intercept !== 0.0)
   }
 
   test("logistic regression doesn't fit intercept when fitIntercept is off") {
     val lr = new LogisticRegression
     lr.setFitIntercept(false)
     val model = lr.fit(dataset)
-    assert(model.intercept == 0.0)
+    assert(model.intercept === 0.0)
   }
 
   test("logistic regression with setters") {