From: Kenny Tilton
Subject: [OT][Long][You All Know Everything Dept] Recommendations for Stat book?
Date: 
Message-ID: <YK%xc.172576$WA4.15245@twister.nyc.rr.com>
This OT post falls in the "Lispniks Know Everything Department". The 
last time I mentioned that I got a nice borscht recipe. Now someone is 
looking for a good stat book, and reported no luck with NGs specific to 
that. Here is a doozy of a spec:

"I am looking for a statistics and probability textbook that would be 
small in size, but pithy in
explanation.

I below list a few books that I have looked at and the deficiencies that 
I found with them:

Example 1)
"Introduction to Probability and Statistics", Mendenhall, Beaver, 
Beaver, 10th edition:

http://www.amazon.com/exec/obidos/tg/detail/-/0534357784/qid=1086880204/sr=1-3/ref=sr_1_3/002-3751143-0849664?v=glance&s=books

I found it to be too light in terms of content. For example, it defined 
in 2nd chapter what
population variance and sample variance is, and that in case of one you 
divide by n and in case of
the other by n-1. Later in the text, and I quote:

"You may wonder why you need to divide by (n-1) rather than n when 
computing sample variance ...
turns out that the sample variance s^2 with (n-1) in the denominator 
provides better estimates of
(sigma)^2 than would an estimator calculated with n in the denominator"

And that's it. As a mathematically curious person with college calculus 
experience, I find that to
be extremely intellectually unsatisfying. I _COULD_ google for the 
missing information, but having
it all in one source would be heavenly. Having thumbed through the rest 
of the book, I observed
that pretty much nothing is formally proven or explained. The book comes 
in at 750+ pages and is
physically a clunker, which is unjustifiable for the amount of 
information it omits. There is an
inordinate amount of time spent on how to use the included Minitab, with 
screenshots and all. I
need a textbook, not a software manual.

Example 2)

"Probability and Statistics for Engineering and the Sciences" by Jay L. 
Devore 5th edition

http://www.amazon.com/exec/obidos/tg/detail/-/0534372813/002-3751143-0849664?v=glance

Here is an example of the sort of content that I am looking for. 
Everything is formally derived,
using univariate calculus when necessary.

Problems:

the book is extremely slow reading, but not due to the included proofs. 
In the introduction, the
author mentions how he spent an inordinate amount of time researching 
"real-life examples" because
he found that students are more interested and motivated to learn the 
subject if they are
presented with something other than "artificial examples with little 
variation". I would rather he
had omitted all of that, and made the text a faster read. And the 
size/weight of the book is once
again an issue just like the size of the book in example 1.

Most of the learning that I will be doing of this is during my hour-long 
commute on the New York
City subway, standing with a laptop and a few other books in my 
backpack. Something physically
small and light would be really nice to have."

Me, I think the guy should stick to Harry Potter on the subway, but if 
anyone can think of a suitable title I will relay it back.

kenny

-- 
Home? http://tilton-technology.com
Cells? http://www.common-lisp.net/project/cells/
Cello? http://www.common-lisp.net/project/cello/
Why Lisp? http://alu.cliki.net/RtL%20Highlight%20Film
Your Project Here! http://alu.cliki.net/Industry%20Application

From: Ari Johnson
Subject: Re: [OT][Long][You All Know Everything Dept] Recommendations for Stat book?
Date: 
Message-ID: <Lt0yc.9949$E03.1085@fed1read01>
Kenny Tilton wrote:
> Me, I think the guy should stick to Harry Potter on the subway, but if 
> anyone can think of a suitable title I will relay it back.

And me, I think that even if I could remember the title of the one I 
really liked in my college stats class, I'd still do him a favor and 
give a borscht recipe! :D

The book below is the one the professor I had is using for the same 
class this fall.  I can't guarantee it's the same one I had, but it 
seems likely.  If it is, it was a darn good book that I'm almost sad I 
sold back to the bookstore.

AN INTRODUCTION TO MATHEMATICAL STATISTICS and Its Applications 3rd 2001 
by Larson/Marx; Publ: Prentice Hall; ISBN: 0-13-922303-7
From: Erann Gat
Subject: Re: [OT][Long][You All Know Everything Dept] Recommendations for Stat book?
Date: 
Message-ID: <gNOSPAMat-221501.10150410062004@nntp1.jpl.nasa.gov>
In article <······················@twister.nyc.rr.com>,
 Kenny Tilton <·······@nyc.rr.com> wrote:

> And the size/weight of the book is once
> again an issue just like the size of the book in example 1.

Try "Introduction to Probability Theory" by Hoel, Port and Stone.  It's 
the first in a three-volume series.  It doesn't do any handwaving (as 
far as I can recall) and it has a nice compact form factor.

I haven't looked at the other two books in the series (which cover 
statistics and stochastic processes) but if they are similar to the 
first one then they are probably also worthwhile.

E.
From: Joe Marshall
Subject: Re: [OT][Long][You All Know Everything Dept] Recommendations for Stat book?
Date: 
Message-ID: <8yevte3v.fsf@ccs.neu.edu>
Kenny Tilton <·······@nyc.rr.com> writes:

> This OT post falls in the "Lispniks Know Everything Department". The
> last time I mentioned that I got a nice borscht recipe. Now someone is
> looking for a good stat book, and reported no luck with NGs specific
> to that. 

For an `unorthodox' approach, try
   http://omega.math.albany.edu:8008/JaynesBook.html

Jaynes was an extremely bright physicist who rejected standard
probability and statistics in favor of the Bayesian approach.  You
might mistake Jaynes for a crackpot because he's outside the
mainstream of statistics, but his approach is solidly grounded and
gives better answers.
From: Christophe Rhodes
Subject: Re: [OT][Long][You All Know Everything Dept] Recommendations for Stat book?
Date: 
Message-ID: <sqbrjrtab3.fsf@cam.ac.uk>
Joe Marshall <···@ccs.neu.edu> writes:

> Kenny Tilton <·······@nyc.rr.com> writes:
>
>> This OT post falls in the "Lispniks Know Everything Department". The
>> last time I mentioned that I got a nice borscht recipe. Now someone is
>> looking for a good stat book, and reported no luck with NGs specific
>> to that. 
>
> For an `unorthodox' approach, try
>    http://omega.math.albany.edu:8008/JaynesBook.html
>
> Jaynes was an extremely bright physicist who rejected standard
> probability and statistics in favor of the Bayesian approach.  You
> might mistake Jaynes for a crackpot because he's outside the
> mainstream of statistics, but his approach is solidly grounded and
> gives better answers.

Another Bayesian-oriented book, though not primarily about statistics,
is MacKay's Information Theory, Inference and Learning Algorithms.
It's based around a course on Information Theory and Pattern
Recognition for physicists; the mathematical content is not terribly
tricky, but it does discuss uses of statistics (both sampling theory
and Bayesian inference) in detail.  Very good on communication over
noisy channels, too.

Christophe
-- 
http://www-jcsu.jesus.cam.ac.uk/~csr21/       +44 1223 510 299/+44 7729 383 757
(set-pprint-dispatch 'number (lambda (s o) (declare (special b)) (format s b)))
(defvar b "~&Just another Lisp hacker~%")    (pprint #36rJesusCollegeCambridge)
From: Kenny Tilton
Subject: Re: [OT][Long][You All Know Everything Dept] Recommendations for Stat book?
Date: 
Message-ID: <zb4yc.173116$WA4.136662@twister.nyc.rr.com>
Thx, Christophe and everyone, I am forwarding these as they come in. dk.

Christophe Rhodes wrote:
> Joe Marshall <···@ccs.neu.edu> writes:
> 
> 
>>Kenny Tilton <·······@nyc.rr.com> writes:
>>
>>
>>>This OT post falls in the "Lispniks Know Everything Department". The
>>>last time I mentioned that I got a nice borscht recipe. Now someone is
>>>looking for a good stat book, and reported no luck with NGs specific
>>>to that. 
>>
>>For an `unorthodox' approach, try
>>   http://omega.math.albany.edu:8008/JaynesBook.html
>>
>>Jaynes was an extremely bright physicist who rejected standard
>>probability and statistics in favor of the Bayesian approach.  You
>>might mistake Jaynes for a crackpot because he's outside the
>>mainstream of statistics, but his approach is solidly grounded and
>>gives better answers.
> 
> 
> Another Bayesian-oriented book, though not primarily about statistics,
> is MacKay's Information Theory, Inference and Learning Algorithms.
> It's based around a course on Information Theory and Pattern
> Recognition for physicists; the mathematical content is not terribly
> tricky, but it does discuss uses of statistics (both sampling theory
> and Bayesian inference) in detail.  Very good on communication over
> noisy channels, too.
> 
> Christophe

-- 
Home? http://tilton-technology.com
Cells? http://www.common-lisp.net/project/cells/
Cello? http://www.common-lisp.net/project/cello/
Why Lisp? http://alu.cliki.net/RtL%20Highlight%20Film
Your Project Here! http://alu.cliki.net/Industry%20Application
From: Gareth McCaughan
Subject: Re: [OT][Long][You All Know Everything Dept] Recommendations for Stat book?
Date: 
Message-ID: <87fz93c8yv.fsf@g.mccaughan.ntlworld.com>
Christophe Rhodes <·····@cam.ac.uk> writes:

> Joe Marshall <···@ccs.neu.edu> writes:
> 
> > Kenny Tilton <·······@nyc.rr.com> writes:
> >
> >> This OT post falls in the "Lispniks Know Everything Department". The
> >> last time I mentioned that I got a nice borscht recipe. Now someone is
> >> looking for a good stat book, and reported no luck with NGs specific
> >> to that. 
> >
> > For an `unorthodox' approach, try
> >    http://omega.math.albany.edu:8008/JaynesBook.html
> >
> > Jaynes was an extremely bright physicist who rejected standard
> > probability and statistics in favor of the Bayesian approach.  You
> > might mistake Jaynes for a crackpot because he's outside the
> > mainstream of statistics, but his approach is solidly grounded and
> > gives better answers.
> 
> Another Bayesian-oriented book, though not primarily about statistics,
> is MacKay's Information Theory, Inference and Learning Algorithms.
> It's based around a course on Information Theory and Pattern
> Recognition for physicists; the mathematical content is not terribly
> tricky, but it does discuss uses of statistics (both sampling theory
> and Bayesian inference) in detail.  Very good on communication over
> noisy channels, too.

Jaynes's book is available in dead-tree form too. Caveats about it:
  - it's rather long
  - it's sometimes quite technical
  - it's unfinished: Jaynes died before he could finish it,
    and the editor who took it over decided not to try to
    write extra bits for him
  - it's a bit over-polemical in spots (not entirely
    without reason, and quite entertainingly, but still...)

MacKay's is very fine, and available in PDF for free on the
web. But
  - it's also somewhat bulkier than your friend is probably
    looking for
  - it covers a lot of material other than probability and
    statistics

I have a few other decent books on probability and stats
on my bookshelves, but I think they're all rather too
mathematically hairy; they tend to begin along the lines
of
    Let S be any set. A sigma-algebra on S is a set of
    subsets of S such that [...]. If F is a sigma-algebra
    on S, then a probability measure on F is a function
    from F to [0,1] such that [...]. A probability space
    is a tuple <S,F,P> where S is a set, F is a sigma-
    -algebra on S, and P is a probability measure on F.

They also concentrate on probability rather than statistics;
I don't think any of them would bother to explain the point
of the (n-1)/n correction to the sample variance, for
instance. For reference, the books in question are
Grimmett & Stirzaker, "Probability and random processes";
Feller, "An introduction to probability theory and its
applications, vol. I" (much less cuddly than the title
suggests, though it *is* deservedly a classic), and
Williams, "Probability with martingales".

The nearest thing I have to what your friend is after is
M J Moroney's classic (from the 1960s or 1970s, IIRC)
"Facts from Figures". It's intended for the layperson,
so there's not much rigorous proof there, but it's short;
it covers a reasonable amount of ground; it focuses on
statistics rather than probability; it's well written.
Unfortunately it's also out of print. It can be had very
cheaply second-hand via Amazon, and doubtless elsewhere
too.

-- 
Gareth McCaughan
.sig under construc
From: thomas
Subject: Re: [OT][Long][You All Know Everything Dept] Recommendations for Stat book?
Date: 
Message-ID: <3c5586ca.0406110402.5be2bb7c@posting.google.com>
Joe Marshall <···@ccs.neu.edu> wrote in message news:<············@ccs.neu.edu>...
> Kenny Tilton <·······@nyc.rr.com> writes:
> 
> > This OT post falls in the "Lispniks Know Everything Department". The
> > last time I mentioned that I got a nice borscht recipe. Now someone is
> > looking for a good stat book, and reported no luck with NGs specific
> > to that. 
> 
> For an `unorthodox' approach, try
>    http://omega.math.albany.edu:8008/JaynesBook.html
> 
> Jaynes was an extremely bright physicist who rejected standard
> probability and statistics in favor of the Bayesian approach.  You
> might mistake Jaynes for a crackpot because he's outside the
> mainstream of statistics, but his approach is solidly grounded and
> gives better answers.

In the same area (on Jaynes' recommended reading list), but much
slimmer and cheaper in dead-tree, i rather liked:

Data Analysis: A Bayesian Tutorial
D.S. Sivia
ISBN: 0198518897 

It seemed to provide a decent level of 'pick up and use' while taking
care to provide sufficient explanation & motivation.

thomas
From: Erann Gat
Subject: Re: [OT][Long][You All Know Everything Dept] Recommendations for Stat book?
Date: 
Message-ID: <gNOSPAMat-189AAC.12112611062004@nntp1.jpl.nasa.gov>
In article <············@ccs.neu.edu>, Joe Marshall <···@ccs.neu.edu> 
wrote:

> Kenny Tilton <·······@nyc.rr.com> writes:
> 
> > This OT post falls in the "Lispniks Know Everything Department". The
> > last time I mentioned that I got a nice borscht recipe. Now someone is
> > looking for a good stat book, and reported no luck with NGs specific
> > to that. 
> 
> For an `unorthodox' approach, try
>    http://omega.math.albany.edu:8008/JaynesBook.html
> 
> Jaynes was an extremely bright physicist who rejected standard
> probability and statistics in favor of the Bayesian approach.  You
> might mistake Jaynes for a crackpot because he's outside the
> mainstream of statistics, but his approach is solidly grounded and
> gives better answers.

I second this.  I've read the first few chapters of the Jaynes book and 
I think it is extraordinary.  I highly recommend taking a look.

E.
From: Edi Weitz
Subject: Re: [OT][Long][You All Know Everything Dept] Recommendations for Stat book?
Date: 
Message-ID: <87u0xjbalk.fsf@bird.agharta.de>
On Thu, 10 Jun 2004 16:11:36 GMT, Kenny Tilton <·······@nyc.rr.com> wrote:

> Me, I think the guy should stick to Harry Potter on the subway, but
> if anyone can think of a suitable title I will relay it back.

"Cartoon Guide to Statistics" (only half-joking):

  <http://www.amazon.com/exec/obidos/tg/detail/-/0062731025/>

Edi.
From: Raymond Toy
Subject: Re: [OT][Long][You All Know Everything Dept] Recommendations for Stat book?
Date: 
Message-ID: <sxd7juf15dw.fsf@edgedsp4.rtp.ericsson.se>
>>>>> "Edi" == Edi Weitz <···@agharta.de> writes:

    Edi> On Thu, 10 Jun 2004 16:11:36 GMT, Kenny Tilton <·······@nyc.rr.com> wrote:
    >> Me, I think the guy should stick to Harry Potter on the subway, but
    >> if anyone can think of a suitable title I will relay it back.

    Edi> "Cartoon Guide to Statistics" (only half-joking):

    Edi>   <http://www.amazon.com/exec/obidos/tg/detail/-/0062731025/>

Haven't read this, but his Cartoon History of the Universe 1 and 2 are
quite fun.

Ray
From: R. Scott McIntire
Subject: Re: [OT][Long][You All Know Everything Dept] Recommendations for Stat book?
Date: 
Message-ID: <8a9yc.25089$HG.6078@attbi_s53>
"Kenny Tilton" <·······@nyc.rr.com> wrote in message
···························@twister.nyc.rr.com...
>
>
http://www.amazon.com/exec/obidos/tg/detail/-/0534357784/qid=1086880204/sr=1-3/ref=sr_1_3/002-3751143-0849664?v=glance&s=books
>
> I found it to be too light in terms of content. For example, it defined
> in 2nd chapter what
> population variance and sample variance is, and that in case of one you
> divide by n and in case of
> the other by n-1. Later in the text, and I quote:
>
> "You may wonder why you need to divide by (n-1) rather than n when
> computing sample variance ...
> turns out that the sample variance s^2 with (n-1) in the denominator
> provides better estimates of
> (sigma)^2 than would an estimator calculated with n in the denominator"
>
> And that's it. As a mathematically curious person with college calculus
> experience, I find that to
> be extremely intellectually unsatisfying.

I usually find that you get books that a recipies or a theorem definition
parade, when what you would like is some intuition about the subject backed
up with a reaonable explaination. Let me explain a little bit about sample
variance.

It turns out that if I you n identically distributed random variables x_i
{i=1,..n} with mean m and standard variance s^2, then people combine these
in order to produce an estimator for the mean or the standard deviation. If
we take the n random variables and form a NEW RANDOM VARIABLE x_bar that is
the average value of the other random variables then this new random
variable can be used as an estimator of the mean m. It is also an UNBIASED
estimator as the expected value of this random variable is m. You can do the
same thing to get an estimator of the variance s^2. The random variable
s_bar^2 = (sum (x_i - m)^2 )/ n. You can check that the expectation of this
random variable is s^2.

is an unbiased estimator of the variance of the random variables. That is,
the expected value of this random variable is s^2. Now, although we know
that
all the random variables have the same mean, m, and variance, s^2, suppose
we don't know what they are. Well we didn't need to know them for the first
estimator - for the mean - ok. But, for the variance estimator, we need to
know m. If we don't have it, we could try to use the estimator for m. This
gives a new estimator for the variance:
s_2^2 = (sum x_i - x_bar)^2 / n.

But now this estimator is unbiased. The expectation of this random variable
is not s^2. A tedious calculation will show that you need to divide by n-1
rather than n to make it unbiased. Finally, in the real world we replace the
x_i's with actual sample data that is suppose to come from identically
distributed random variables.

Hope this helps.

-R. Scott McIntire
From: Kaz Kylheku
Subject: Re: [OT][Long][You All Know Everything Dept] Recommendations for Stat book?
Date: 
Message-ID: <cf333042.0406161821.6375cfc3@posting.google.com>
Kenny Tilton <·······@nyc.rr.com> wrote in message news:<······················@twister.nyc.rr.com>...
> "I am looking for a statistics and probability textbook that would be 
> small in size, but pithy in
> explanation.

I have a statistics book which, based on random sampling of statistics
books, it is merely in the fifth percentile among them in size, with
19 out of 20 confidence.
From: T.M. Sommers
Subject: Re: [OT][Long][You All Know Everything Dept] Recommendations for Stat book?
Date: 
Message-ID: <ANmdnRofpK-XXErdUSdV9g@telcove.net>
Kenny Tilton wrote:
> 
> "You may wonder why you need to divide by (n-1) rather than n when 
> computing sample variance ...
> turns out that the sample variance s^2 with (n-1) in the denominator 
> provides better estimates of
> (sigma)^2 than would an estimator calculated with n in the denominator"
> 
> And that's it. As a mathematically curious person with college calculus 
> experience, I find that to
> be extremely intellectually unsatisfying. 

I like what the authors of _Numerical Recipes_ have to say on 
this issue: "We might also comment that if the difference between 
N and N-1 ever matters to you, then you are probably up to no 
good ... ."

> Me, I think the guy should stick to Harry Potter on the subway, but if 
> anyone can think of a suitable title I will relay it back.

Harry Potter and the Analysis of Variance?


-- 
Thomas M. Sommers -- ···@nj.net -- AB2SB