[SPARK-10959] [PYSPARK] StreamingLogisticRegressionWithSGD does not train with given regParam and convergenceTol parameters

These params were being passed into the StreamingLogisticRegressionWithSGD constructor, but not transferred to the call for model training.  Same with StreamingLinearRegressionWithSGD.  I added the params as named arguments to the call and also fixed the intercept parameter, which was being passed as regularization value.

Author: Bryan Cutler <[email protected]>

Closes apache#9002 from BryanCutler/StreamingSGD-convergenceTol-bug-10959.

(cherry picked from commit 5410747)
Signed-off-by: Xiangrui Meng <[email protected]>

Author: BryanCutler

python/pyspark/mllib/classification.py

@@ -632,7 +632,8 @@ def update(rdd):
             if not rdd.isEmpty():
                 self._model = LogisticRegressionWithSGD.train(
                     rdd, self.numIterations, self.stepSize,
-                    self.miniBatchFraction, self._model.weights)
+                    self.miniBatchFraction, self._model.weights,
+                    regParam=self.regParam, convergenceTol=self.convergenceTol)
 
         dstream.foreachRDD(update)
 

python/pyspark/mllib/regression.py

@@ -669,7 +669,7 @@ def update(rdd):
                 self._model = LinearRegressionWithSGD.train(
                     rdd, self.numIterations, self.stepSize,
                     self.miniBatchFraction, self._model.weights,
-                    self._model.intercept)
+                    intercept=self._model.intercept, convergenceTol=self.convergenceTol)
 
         dstream.foreachRDD(update)