[MRG] Consistent prediction results for simple_differencing = True and False

[drxyzw]

Description

SARIMA and ARIMA predict() and predict_in_sample() now return undifferenced prediction and confidence intervals if simple_differencing = True. When d or D is not 0, the two functions used to return differenced results. This was not consistent with behaviors of simple_differencing = False and corresponding R library.

Fixes #421

Type of change

• [x] Bug fix (non-breaking change which fixes an issue)

How Has This Been Tested?

Added two tests to pmdarima\pmdarima\arima\tests\test_auto.py. The tests to check if the prediction results are similar for simple_differencing = True and False.

• [x] test_stepwise_with_simple_differencing2()
• [x] test_stepwise_with_simple_differencing3()

Checklist:

• [x] I have performed a self-review of my own code
• [x] I have commented my code, particularly in hard-to-understand areas
• [ ] I have made corresponding changes to the documentation
• [x] My changes generate no new warnings
• [x] I have added tests that prove my fix is effective or that my feature works
• [x] New and existing unit tests pass locally with my changes

[tgsmith61591]

@drxyzw this is a great PR, thanks so much for your contribution, and sorry for the time it took me to get back to you. I've been taking some time off.

[drxyzw]

No problem, and thank you for your review and approval.