Inefficient use of XtX calculation in Ordinary Least Squares

In line 59 of OrdinaryLeastSquaresModel we calculate and collect XtX, but then don't bother to use it (so it is in essence calculated twice).

~https://github.com/apache/zeppelin/blob/master/spark/src/main/java/org/apache/zeppelin/spark/SparkInterpreter.java~

Not sure how that link got in there. It is unrelated.