Refactor StackingCVRegressor so that users can obtain meta_features

[takashioya]

Description

modify fit and add an attribute self.train_meta_features_ for obtaining train meta-features,
and add a method predict_meta_features for obtaining test meta-features.
Sorry, I closed the previous PR #294 by mistake, so I opened new one.

Related issues or pull requests

#285
#290
#294

Pull Request requirements

• Added appropriate unit test functions in the ./mlxtend/*/tests directories
• Ran nosetests ./mlxtend -sv and make sure that all unit tests pass
• Checked the test coverage by running nosetests ./mlxtend --with-coverage
• Checked for style issues by running flake8 ./mlxtend
• Added a note about the modification or contribution to the ./docs/sources/CHANGELOG.md file
• Modify documentation in the appropriate location under mlxtend/docs/sources/ (optional)
• Checked that the Travis-CI build passed at https://travis-ci.org/rasbt/mlxtend

[coveralls]

Coverage increased (+0.02%) to 90.897% when pulling e1cbe25 on takashioya:refactor-meta-features into cca77ac on rasbt:master.

[rasbt]

Thanks for making those changes. The PR looks good to me now :). Unless you have further comments, I am happy to merge this (and I will take care of the API documentation update for the website upon the next version release)

[takashioya]

No comments, thanks for the good advice for those changes!
'Python Machine Learning'(in Japanese version) is my favorite book and I have thought I'd like to contribute to your library some time, so I'm happy :)

[rasbt]

Nice, will merge then, thanks again! :) And I am also happy to hear that you like the book :)