QMCD normalization fix

[KulikDM]

All Submissions Basics:

• Have you followed the guidelines in our Contributing document?
• Have you checked to ensure there aren't other open Pull Requests for the same update/change?
• Have you checked all Issues to tie the PR to a specific one?

All Submissions Cores:

• Have you added an explanation of what your changes do and why you'd like us to include them?
• Have you successfully ran tests with your changes locally?
• Does your submission pass tests, including CircleCI, Travis CI, and AppVeyor?
• Does your submission have appropriate code coverage? The cutoff threshold is 95% by Coversall.

Hi, this PR addresses the Issue #549
All the previous functionality still exists.
Notable changes include: returned decision scores are no longer normalized, better criterion for switching the scores when necessary, exposing the decision_function to correctly handle new data not seen during the OD fitting.

Note that the last change above will cause the model to hold more memory as the original dataset must be stored as variable with which to compare the new data to. This will affect a stored joblib model size if created.

If there are any issues let me know and am happy to make the necessary changes.

[coveralls]

Pull Request Test Coverage Report for Build 8786241040

Details

• 17 of 21 (80.95%) changed or added relevant lines in 1 file are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage decreased (-0.009%) to 93.744%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
pyod/models/qmcd.py	17	21	80.95%

Totals
Change from base Build 8309419980:	-0.009%
Covered Lines:	9006
Relevant Lines:	9607

---

💛 - Coveralls