sci.stat.math Google Group

Re: the distribution of the ratio of two binomial random variables?

2009-11-29T03:35:20Z

The variance is approximately p1*p2*(p1*(1-p2)+(1-p1)*p2)/(n (p1+p2)^4)
(Think "beta".)

cheap louis vuitton coach supra boss cole haan shoes,cheap nike shoes jordan shoes jordan AF1 fusion shoes

2009-11-29T03:09:48Z

( [link] ) China Supply ED Hardy,Versace Women
Sandals,Boots
( [link] ) Cheap Wholesale Coach,Chanel Lady
Slippers,Sandals
( [link] ) Puma,Christian Louboutin Sandals
( [link] ) Discount Gucci Prada Sneakers First
Quality,Classic

Re: the distribution of the ratio of two binomial random variables?

2009-11-28T19:24:22Z

Rationale
The Reader perfreem is concerned with the Expectation of a certain function of Binomial random variables, namely Z=X/(X+Y). Retaining the particular case, as it was asked for, that n is a fixed constant: X=Bin(pX, n), Y=Bin (pY, n) we did evaluate E(Z) considering:
_____Z=X/(X+Y)________0<= X <=n____0<=Y<=n

Re: Correlation and Observations with Zero Value

2009-11-28T17:19:52Z

Yes, I saw it.
I was just trying to understand why Graham was worried about
zeros, and speculated that he might be concerned about it if 0
represented a lower boundary. I guess I was also thinking about
the possibility of cases like the 4th Anscombe (see link below),
but with the bunch of points falling at 0.

HOT!!! 2009 Cheap wholesale True Relig Jeans at www.fjrjtrade.com

2009-11-28T15:52:00Z

HOT!!! 2009 Cheap wholesale True Relig Jeans at [link]
<paypal payment>
Cheap Wholesale Jeans [link]
Cheap Wholesale True Relig Jeans [link]
Cheap Wholesale True Relig Jeans [link]
Cheap Wholesale others Jeans [link]
╃Men Size 30,32,34,36,38,40 Women Size 26,27,28,29,30,31 Cheap

(Paypal Payment)Hot sale Fashion Nike Jordan,shox,free,rift,max,AF1,Bape,Gucci,Prada, shoes Wholesale and retail Welcome to www.ebaychinaonline.com

2009-11-28T11:39:05Z

For more information,please contact Sophie,Yahoo:
globaltrader2...@yahoo.com.cn
Minimum order is one,factory price also!Paypal payment free
shipping,ship time will take 4-7 working days.
wholesale Nike Air Jordan 1 Seller Paypal Payment
[link]
Discount Nike Air Jordan 2 Shoes Seller Paypal Payment

Re: Correlation and Observations with Zero Value

2009-11-28T00:02:25Z

On Thu, 26 Nov 2009 21:21:54 EST, Graham Ashe
I'm wondering about several things.
What is your correlation *for*? Why do you *imagine*
that you should assume it is zero?
I'm wondering if my post yesteday got through. I see it on my reader.
I'm wondering why Greg and Bruce are concerned about the number

Re: prior information in mean and variance in regression

2009-11-27T23:43:41Z

On Thu, 26 Nov 2009 12:58:11 -0800, Richard Startz
Bruce, Dick,
I was thinking of what Dick wrote when I wrote my post -
I thought that I knew what he was intending, and I tried
to sharpen it a little. But I didn't do a great job of saying
it, either. Dick does somewhat better, above.
Here is another attempt.

Simulating fractiles N(0,1)

2009-11-27T22:30:53Z

Table of the ABSOLUTE differences between the empirical fractiles (x10^6) and theoretical cumulative values [*]. Data N(0,1), total simulations of each attempt = 4E6, 5 attempts.
****************************** ******************
______95%____97.5%____99%____9 9.9%____
____1.64485__1.95996__2.32635_ _2.57583_________[*]

Re: the distribution of the ratio of two binomial random variables?

2009-11-27T19:15:17Z

p1/(p1+p2) seems to approximate the mean.

Free shipping Cheap Wholesale LV Handbag, Jimmy Choo,Miumiu Real Leather Handbags&purse (www.dotradenow.com.cn)

2009-11-27T18:49:00Z

AAA real leather Handbags [link]
Cheap Wholesale Balenciaga real leather Handbags
<[link] paypal payment>
Cheap Wholesale Balenciaga real leather Purse
Cheap Wholesale Bally real leather Purse <free shipping paypal
payment>
Cheap Wholesale BOSS real leather Purse <[link] paypal

Hot Sale!!! 2009 High quality cheap wholesale LV Shoes at www.fjrjtrade.com

2009-11-27T18:24:45Z

Hot Sale!!! 2009 High quality cheap wholesale LV Shoes at [link]
<Paypal Payment>
Welcome to visit [link]
Men size 40,41,42,43,44,45,46. Women size 36,37,38,39,40.
High quality wholesale Air Force One shoes, Nike Jordan, Nike,Air Max,
Nike Shox, UGG Shoes, Puma Shoes, Nike shoes, Adidas Shoes, Christian

Re: Is Huber Estimator IRLS Weight Wrong in Literature?

2009-11-27T17:30:22Z

Bump. Are you geniuses dumbfounded?

Re: Correlation and Observations with Zero Value

2009-11-27T17:00:07Z

Why?
Not "all". You said negative values are possible.
Please put your problem in context.
What do the two variables represent?
Is one a predictor and the other a response or are both predictors?
Are the zeros predictor values or response values?
Why are negative values so rare?
Hope this helps.
Greg

Re: calculation of fishers exact test

2009-11-27T15:34:17Z

I have a 4x2 contingency table and have the same problem. Did you get your problem solved?
Could you send me the recipe?
Thanks
Alex

Discount Wholesale A&F Pants Affliction Jeans Coogi Jeans D&G ED Hardy G-STAR True Religion Jeans Pants Belts etc ==>

2009-11-27T14:55:09Z

Discount Wholesale Affliction Jeans <free shipping paypal payment>
Discount Wholesale AK Jeans ( [link] )
Discount Wholesale Armani Jeans
Discount Wholesale Artful Dodger Jeans <free shipping paypal payment>
Discount Wholesale BAPE Jeans
Discount Wholesale BBC Jeans ( [link] )

www.ebaychinaonline.com Wholesale prada,bape,converse,ugg shoes,ugg spring shoes,ugg boots,air jordan,puma shoes at shock price,factory promotion,paypal payment

2009-11-27T12:17:15Z

For more information,please contact Sophie,Yahoo:
globaltrader2...@yahoo.com.cn
Minimum order is one,factory price also!Paypal payment free
shipping,ship time will take 4-7 working days.
ugg shoes,ugg clog shoes male,UGG shoes women,UGG high heel boots,UGG
low heel shoes,UGG sandals,UGG slippers,UGG spring shoes,UGG kid's

Re: Correlation and Observations with Zero Value

2009-11-27T12:09:22Z

I think you need to look at the scatter-plot, and ask yourself if a
straight line provides a good summary of the relationship between X
and Y. And you need to watch out for the situation seen in the 4th
graph here:
[link]

Cohen's vs. Fleiss Kappa

2009-11-27T07:58:56Z

Hi all, I have a VERY general question: I know about the basic
distinction between Cohen's and Fleiss kappa (one is for TWO raters
and the other is for more than two..). Going a bit beyond that: what's
the trick (math) behind each one test??
I'm curious about it because when I calculate Fleiss' kappa for data

Re: Mechanical generation of random bit sequences

2009-11-27T07:03:35Z

gettimeofday() and whatever goes into select(), will both typically be
driven off the interval timer, and will not resolve times less than
whatever that's set to. In some environments, that resolution can
change (Windows’s timeBeginPeriod(), for example).
Agreed. Using all (not just keyboard) interrupts, plus the local

Re: Correlation and Observations with Zero Value

2009-11-27T02:21:54Z

Actually, no. I guess I misunderstood him to mean that '0' was, in theory, a perfectly acceptable value for one of my observations.

'0' is pretty much the lower limit (negative values for observations are possible but very rare). What I'm concerned about is whether too many (say, more than 50%) '0'-value observations in the sample affects the validity or accuracy of correlations calculated using that sample. I mean, would it be better to simply dismiss the resulting correlation coefficient as '0' itself (i.e. neither positive nor negative)?

www.urban4wholesaler.com discount cheap ed hardy gucci coogi lacoste polo urban t-shirts,cheap fendi coach juicy miumiu chloe handbags wallets cheap airmax ltd tn 90 shoes

2009-11-27T02:05:08Z

Cheap Wholesale A&F T-Shirt
Cheap Wholesale Adidas T-Shirt low price with high quality
( [link] )
Cheap Wholesale AFF T-Shirt low price with high quality
Cheap Wholesale Akademiks T-Shirt ( [link] )
Cheap Wholesale Armani T-Shirt
Cheap Wholesale Baby T-Shirt low price with high quality

Re: Mechanical generation of random bit sequences

2009-11-26T22:19:56Z

I'm afraid you're right. While using the RDTSC works good, the
hardware in the keyboard and in the ports (PS2, USB) and probably the
software make the resolution much worse. You loose most of the
information this way and maybe get some entropy from the jitter
between the CPU clock and the keyboard clock (or whatever).

Re: Correlation and Observations with Zero Value

2009-11-26T21:12:56Z

Yep.
Me too.
Greg

Re: regression question

2009-11-26T21:05:47Z

The number of variables and their values would have to be orders of
magnitude higher before this would be a concern. Re the latter, I
always standardize non=binary variables.
Hope this helps.
Greg

Re: prior information in mean and variance in regression

2009-11-26T20:58:11Z

On Thu, 26 Nov 2009 11:42:44 -0800 (PST), Bruce Weaver
Bruce:
Let me try to be more clear. The regression models the CONDITIONAL
mean of y, which depends on X. The intercept isn't the mean of y and
the standard error of the regression estimates the standard deviation
of the error term, not of y. That's what I meant by saying the mean

Re: Expected value involving max() function and Normal R.V.

2009-11-26T20:52:30Z

If you are going to post in multiple groups.Send one post
to a comma separated list instead of a separate post for
each group.
Greg

Re: Negative R-squared value?

2009-11-26T20:09:55Z

There are two conventions in play for "Total Sum of Squares" and
thus for R^2, and you need to keep them distinct and clear.
TSS_ols is always defined for OLS-with-an-intercept as being
the "mean-adjusted total sum of squares" -- the sum of squares,
taken around the mean.
TSS_nl for non-linear regression is *usually* defined as being

Re: Correlation and Observations with Zero Value

2009-11-26T19:54:17Z

On Wed, 25 Nov 2009 23:05:04 EST, Graham Ashe
You are concerned with "too many zeroes" - which must be
hard to distinguish from "too many of any single value".
(I don't see why another poster is concerned whether 0 is
a lower limit.) The product-moment correlation remains
well defined for weights at a few points, or at many points.

Re: prior information in mean and variance in regression

2009-11-26T19:42:44Z

Dick, I've been trying for a while to understand what you're getting
at in that first sentence, but so far no luck. I don't understand how
the mean and variance of Y can be "almost irrelevant" when linear
regression entails partitioning the sum of the squared deviations
about the MEAN of Y (i.e., the numerator of the variance) into

Re: prior information in mean and variance in regression

2009-11-26T19:06:32Z

On Tue, 24 Nov 2009 16:59:22 -0800 (PST), soms <soms.sha...@gmail.com>
wrote:
As others have mentioned, the meaning here is unclear.
A "small sample" is improved by larger N drawn randomly from
the same population.
? again. If the "larger sample" does not represent the same
population, then you ought to have great doubts about the validity

Discount Nike jordan Sneakers, Nike all categaries, Nike Shox, Nike Air Max, Air Force One==>==>paypal payment ==

2009-11-26T18:48:01Z

We offer Cheap Nike Shoes, Air Jordan Shoes are on sale now, our Air
Jordan shoes, Nike Shoes are exclusive and we offer original box of
jordan shoes.
Nike Air Jordan Shoes
Cheap Wholesale Jordan 1 (paypal payment)([link] )
Cheap Wholesale Jordan 1.5 (paypal payment)([link] )

Re: Negative R-squared value?

2009-11-26T18:08:27Z

DRG wrote:
One the first question, you are only guaranteed to have RSS <= TSS if you are fitting a model of the form
y= a + b f(x;c)
where a and b are free parameters. The models you quote are not of this form. And it only works if the values of a and b are determined by least squares.
David Jones

METRIC CONVERTER

2009-11-26T17:56:22Z

.
~~~***~~~
============================== ====================
============================== ====================
click here to enter:
>>> [link] <<<
============================== ====================

RESTAURANT REVIEWS

2009-11-26T16:29:04Z

Re: regression question

2009-11-26T16:15:37Z

Thank you, Greg. I downloaded the netlab toolbox, mapped it in MATLAB,
and I am going thorough the learning curve now with the aid of the
help files.
With the significant detail involved with RBF calcs, I am curious if
my 32-bit processor (w/ MS Windows OS) will yield adequate results, or
if I should work on a 128-bit Sun system.

Re: Correlation and Observations with Zero Value

2009-11-26T13:26:48Z

Graham, I understood Greg to be asking if there are negative numbers
in the population. Is that how you read it?
When I read your first post, I wondered if 0 is the minimum possible
score, and if you are therefore worried about a floor effect.

Re: prior information in mean and variance in regression

2009-11-26T15:29:02Z

In
Yi = A*Xi + B + ei
, if you know the value of B, you would be a damn fool not to subtract
B from Yi before using regression
Yi' = A*Xi + ei
where
Yi' = Yi - B
to find A. If E(Xi) = 0 and E(ei) = 0, then B = E(Yi), your "mean".

wholesale soccer jersey(www.jersey-2009.com)

2009-11-26T14:25:27Z

wholesale discounts authentic UGG shoes made in china
([link] )
nike sneakers([link])
AF1 shoes([link])
adidas 35th shoes([link])
nike shoes,jordan shoes([link])
Air Force 1 shoes([link])
kids shoes([link])

NFL nfl jersey Tennessee Titans jerseys www.jersey-2009.com

2009-11-26T14:25:55Z

(paypal payment)( [link]) Nike Air Max 90 Sneakers,
Nike Air Max 91
Sneakers.
(paypal payment)( [link]) Sneakers Supplier Trainers
Wholesale Shop Catalogs Men's Women's Shoes Customize.
(paypal payment)( [link]) Nike Air Max 95 Supplier,
Nike
Air Max 97 Shoes

Re: Negative R-squared value?

2009-11-26T14:09:43Z

Numerical error and negative "sums of squares" is an an old and often repeated
story. The simple example is the use of the text book formaula of summing "x"
and "x squared" rather than summing the squared deviations about the mean. The
problem usually involves a large mean and small deviations. If the information

Re: Expected value involving max() function and Normal R.V.

2009-11-26T13:45:34Z

if I have understood
E[max(0, Q - D)] = integrate (Q-D)*f(D)dD from D= - infinity to Q
Call the answer M =M(Q) which will be some form of erf()
similar thing for the otehr term, the answer is of course m-M
so H(Q) = (a-b)*M(Q) + b*m

Negative R-squared value?

2009-11-26T13:37:32Z

I am trying to test the goodness of fit of some models and have found that for one of the models, I am getting negative R^2 values. I am calculating from R^2 = 1 - RSS/TSS and in the case in question, RSS > TSS. Can anyone shed any light on this?
Secondly, and perhaps more worryingly, the models I am fitting are not strictly linear. For example, one model (which gives good r^2 values) is

Re: Mechanical generation of random bit sequences

2009-11-26T13:20:53Z

I use gettimeofday(&tv_t,NULL); on Linux. I could try ...
struct timespec {
long tv_sec; /* seconds */
long tv_nsec; /* nanoseconds */
};
"Some code calls select() with all three sets empty, nfds zero, and a
non-NULL timeout as a fairly portable way to sleep with subsecond

Expected value involving max() function and Normal R.V.

2009-11-26T11:06:49Z

Hello,
I'm on a deadend road with the following equation:
Eq. 1) H(Q) = a * E[max(0, Q - D)] + b * E[max(0, D-Q)]
where:
- a, b and Q are given constants
- D is a continuous, normally distributed random variable with mean m
and std dev. s
- E[] is the expected value operator
Now, I've found the following result on the web:

≥◇≤Cheap Wholesale Jacket coat fashionable scarfs Sweaters (free shipping www.dotradenow.com.cn)

2009-11-26T09:02:23Z

Suit [link]
Cheap Wholesale Adidas Suit
Cheap Wholesale Baby Suit (wholesale free shipping)
Cheap Wholesale ED Suit
Cheap Wholesale juicy Suit (wholesale free shipping)
Cheap Wholesale Lacoste Suit
Cheap Wholesale NBA Suit
Cheap Wholesale Nike Suit

Re: the distribution of the ratio of two binomial random variables?

2009-11-26T08:59:29Z

Add a third property: Swapping p1 and p2 reverse the pmf
(i.e., reflects it thru Z = 1/2).

Re: the distribution of the ratio of two binomial random variables?

2009-11-26T07:56:45Z

If you exclude (0,0) from the joint distribution of (X,Y), which are
otherwise independent binomials with the same n but possibly different
p's, and let Z = X/(X+Y), then two things are immediately apparent
from looking at a few plots:
1. The values that Z can have are symmetrically placed in [0,1].

Re: Mechanical generation of random bit sequences

2009-11-26T06:55:18Z

A PS/2 style keyboard/mouse UART runs asynchronously from the CPU.
And you're measuring the nanosecond arrival of the keystroke at the
program, which depends on exactly when the key is hit by the user,
when the keystroke is sent by the keyboard, when the keystroke is
received by the UART, when the interrupt is received by the CPU, when

Re: Mechanical generation of random bit sequences

2009-11-26T06:32:33Z

Does it really matter whether the counter on the PC is running at
1GHz, when the keyboard characters on an AT-style keyboard are
clocked in at a frequency more like 20-30KHz? A single character
(scan code) is 11 bits, although some keystrokes are more than one
character. (This may not be an issue on USB keyboards, or maybe

sci.stat.math Google Group

Re: the distribution of the ratio of two binomial random variables?

cheap louis vuitton coach supra boss cole haan shoes,cheap nike shoes jordan shoes jordan AF1 fusion shoes

Re: the distribution of the ratio of two binomial random variables?

Re: Correlation and Observations with Zero Value

HOT!!! 2009 Cheap wholesale True Relig Jeans at www.fjrjtrade.com

(Paypal Payment)Hot sale Fashion Nike Jordan,shox,free,rift,max,AF1,Bape,Gucci,Prada, shoes Wholesale and retail Welcome to www.ebaychinaonline.com

Re: Correlation and Observations with Zero Value

Re: prior information in mean and variance in regression

Simulating fractiles N(0,1)

Re: the distribution of the ratio of two binomial random variables?

Free shipping Cheap Wholesale LV Handbag, Jimmy Choo,Miumiu Real Leather Handbags&purse (www.dotradenow.com.cn)

Hot Sale!!! 2009 High quality cheap wholesale LV Shoes at www.fjrjtrade.com

Re: Is Huber Estimator IRLS Weight Wrong in Literature?

Re: Correlation and Observations with Zero Value

Re: calculation of fishers exact test

Discount Wholesale A&F Pants Affliction Jeans Coogi Jeans D&G ED Hardy G-STAR True Religion Jeans Pants Belts etc ==>

www.ebaychinaonline.com Wholesale prada,bape,converse,ugg shoes,ugg spring shoes,ugg boots,air jordan,puma shoes at shock price,factory promotion,paypal payment

Re: Correlation and Observations with Zero Value

Cohen's vs. Fleiss Kappa

Re: Mechanical generation of random bit sequences

Re: Correlation and Observations with Zero Value

www.urban4wholesaler.com discount cheap ed hardy gucci coogi lacoste polo urban t-shirts,cheap fendi coach juicy miumiu chloe handbags wallets cheap airmax ltd tn 90 shoes

Re: Mechanical generation of random bit sequences

Re: Correlation and Observations with Zero Value

Re: regression question

Re: prior information in mean and variance in regression

Re: Expected value involving max() function and Normal R.V.

Re: Negative R-squared value?

Re: Correlation and Observations with Zero Value

Re: prior information in mean and variance in regression

Re: prior information in mean and variance in regression

Discount Nike jordan Sneakers, Nike all categaries, Nike Shox, Nike Air Max, Air Force One==>==>paypal payment ==

Re: Negative R-squared value?

~~~~~~~~~~~~~~ METRIC CONVERTER ~~~~~~~~~~~~~~

~~~~~~~~~~~~~~ RESTAURANT REVIEWS ~~~~~~~~~~~~~~

Re: regression question

Re: Correlation and Observations with Zero Value

Re: prior information in mean and variance in regression

wholesale soccer jersey(www.jersey-2009.com)

NFL nfl jersey Tennessee Titans jerseys www.jersey-2009.com

Re: Negative R-squared value?

Re: Expected value involving max() function and Normal R.V.

Negative R-squared value?

Re: Mechanical generation of random bit sequences

Expected value involving max() function and Normal R.V.

≥◇≤Cheap Wholesale Jacket coat fashionable scarfs Sweaters (free shipping www.dotradenow.com.cn)

Re: the distribution of the ratio of two binomial random variables?

Re: the distribution of the ratio of two binomial random variables?

Re: Mechanical generation of random bit sequences

Re: Mechanical generation of random bit sequences

METRIC CONVERTER

RESTAURANT REVIEWS