Update classifier.py

[wangg12]

What does this PR do?

To improve the classifier demo by adding an option to train with grid search svm.
Mostly copy paste from the web demo.

Where should the reviewer start?

demos/classifier.py

How should this PR be tested?

test the simple classifier.py demo

Any background context you want to provide?

What are the relevant issues?

[You can link directly to issues by entering # then the
number of the issue, for example, #3 links to issue 3]

Screenshots (if appropriate)

Questions:

• Do the docs need to be updated?
• Does this PR add new (Python) dependencies?
  No
  Add an option to train with grid search svm.

[bamos]

Hi @wangg12 - do you see better performance with the grid search over just a linear SVM with C=1? I used to use the grid search as the default everywhere, but the performance was marginally better than just using a linear SVM and wasn't worth the extra time the grid search took.

I'd be okay adding this since it's just another option. Can you make the following minor changes to your PR:

• Replace print 'training with grid search svm' with the following and make sure your modifications here don't cause flake8/pep8 warnings that will fail the Travis build.

print("""
Warning: In our experiences, using a grid search over SVM hyper-parameters only
gives marginally better performance than a linear SVM with C=1 and
is not worth the extra computations of performing a grid search.
""")

• Import GridSearchCV at the top

[bamos]

Thanks, merged. By the way, I don't get notifications if you update a PR and don't comment on it.

[wangg12]

Thanks, I forgot to comment on it.