Review primitive hyperparams that cause errors

[roquelopez]

There are some default hyperparameters that cause errors every time they are used. For instance, the 'average' hyperparameter of Sklearn Imputer, will always fail for categorical features (we can't calculate the average for this type of feature). These values should be changed, like here.

[madhuripujari95]

The issues seems to be with Countvectorizer and TfidfVectorizer when used with StandardScalar and RobustScalar. (below is the screenshot of the errors)

• Fixed the "ValueError: Cannot center sparse matrices: use with_centering=False instead. See docstring for motivation and alternatives" with the below code in pipeline_builder.py
  elif isinstance(primitive_object, RobustScaler): primitive_object.set_params(with_centering=False)

• Removed standardscalar from grammar and ran the code - all the errors are gone and fasttext on 15-16th rank
  (not sure if this impacts the other pipelines drastically, more testing to be done)