Add drift term keyword to markov.tauchen.

[shizejin]

Try to make markov.tauchen able to handle non-zero drift term in the AR(1) process.

Here discusses the logic of modification made.

[pep8speaks]

Hello @shizejin! Thanks for updating this PR. We checked the lines you've touched for PEP 8 issues, and found:

There are currently no PEP 8 issues detected in this Pull Request. Cheers! 🍻

Comment last updated at 2019-05-23 08:54:28 UTC

[coveralls]

Coverage increased (+0.007%) to 93.95% when pulling 526a891 on shizejin:add_tauchen_drift_term into 26a66c5 on QuantEcon:master.

[jstac]

@shizejin Great work and great analysis.

I was surprised the standard deviation wasn't a bit closer so I tried with m=4 and n=150. It was much better.

(I remember now that the approximation actually gets worse when m gets large.)

@mmcky I approve this PR. I would be glad if we could get this merged and a new version of QE.py released quickly to fix the bug @AnjuJoon found.