Background literature for statsmodels

80 views
Skip to first unread message

Thomas Haslwanter

unread,
Oct 26, 2012, 1:23:59 PM10/26/12
to pystat...@googlegroups.com
Hi,

I am trying to get started in statistical modeling and stats, as I am supposed to hold a stats course at my institution in the next semester.
Having spent some time with S, and having looked at R, I hated the syntax of R, and would like to do as much as possible in Python. "statsmodels" seems to be the leading Python package for statistical modeling. But I have to admit, that with the documentation provided, I have a VERY hard time getting started.

Can anyone recommend any literature on statistical modeling, which might help me getting up to speed?

Thanks,
thomas

josef...@gmail.com

unread,
Oct 26, 2012, 1:55:02 PM10/26/12
to pystat...@googlegroups.com
What field, what topics, what level?

I use mostly econometrics, specialized textbooks and articles.

In statistics, there are some books partially oriented towards R but should
be useful for working with statsmodels.

For many fields there are specifically written textbooks, where statsmodels
might cover some range.

One difference between textbooks and fields is for example the
emphasis on experimental and categorical data versus
non-experimental data and associated problems in econometrics.

Hopefully someone can answer with books that they used.

I can look up some general references but except for a few
econometrics books I never worked through any of them.

The book Skipper and I used most often as reference is
http://pages.stern.nyu.edu/~wgreene/Text/econometricanalysis.htm
1st year PhD level in economics

for econometrics in undergraduate economics, Stock Watson is good
http://www.pearsonhighered.com/stock_watson/

Fox and Weisberg books "sound" good
I never read them, except for some appendices
http://cran.r-project.org/web/packages/car/index.html
http://socserv.mcmaster.ca/jfox/Books/Companion/index.html
http://socserv.socsci.mcmaster.ca/jfox/Books/Applied-Regression-2E/index.html
http://www.stat.umn.edu/alr/

Jonathan's course is useful for statistics
http://www.stanford.edu/class/stats191/


It would be good if we could collect some references for the documentation.
I started a Latex bib file, but didn't get very far yet.

Josef

>
> Thanks,
> thomas

josef...@gmail.com

unread,
Oct 26, 2012, 2:27:07 PM10/26/12
to pystat...@googlegroups.com
------------------
just to have an incomplete list on special topics
mostly advanced topics

time series analysis
~~~~~~~~~~~~~~~

Lütkepohl, used for VAR
http://www.springer.com/economics/econometrics/book/978-3-540-40172-8
Hamilton, the classic
http://press.princeton.edu/titles/5386.html

microeconometrics
~~~~~~~~~~~~~~~
Logit, ... in discrete
Cameron Trivedi
http://cameron.econ.ucdavis.edu/mmabook/mmaprograms.html

robust
~~~~~
Skipper used 1st edition
http://www.amazon.com/Robust-Statistics-Wiley-Probability/dp/0470129905

GLM generalized linear models
~~~~~~~~~~~~~~~~~~~~~~~~
???

others
~~~~~
Owen: empirical likelihood
Li, Racine: kernel density estimation, kernel regression


Stata, SAS, SPSS and R manuals for specific procedures and methods

Josef

>
> Josef
>
>>
>> Thanks,
>> thomas

Skipper Seabold

unread,
Oct 26, 2012, 2:32:35 PM10/26/12
to pystat...@googlegroups.com
Implementation reference and good, basic info

http://www.stata.com/bookstore/generalized-linear-models-and-extensions/

josef...@gmail.com

unread,
Oct 26, 2012, 9:32:03 PM10/26/12
to pystat...@googlegroups.com
(one more)

on the web one of the best collections is by UCLA stats. I look at it
every once in a while
for example
http://www.ats.ucla.edu/stat/dae/
http://www.ats.ucla.edu/stat/examples/default.htm

I wish we had those tutorials for statsmodels, (and had our main gaps filled)

Josef

>>
>> Josef
>>
>>>
>>> Josef
>>>
>>>>
>>>> Thanks,
>>>> thomas

josef...@gmail.com

unread,
Oct 31, 2012, 1:31:04 PM10/31/12
to pystat...@googlegroups.com
(using the thread to add links)

Fundamentals of Modern Statistical Methods
http://www.springer.com/statistics/social+sciences+%26+law/book/978-1-4419-5524-1
looks like a good introductory background with emphasis on robust methods.
(I only skimmed the robust chapter)

Josef
Reply all
Reply to author
Forward
0 new messages