Drexel dragonThe Math ForumDonate to the Math Forum



Search All of the Math Forum:

Views expressed in these public forums are not endorsed by Drexel University or The Math Forum.


Math Forum » Discussions » sci.math.* » sci.stat.math.independent

Topic: The same four proportional weighting factors work for each
00/01/10/11 when 0.25 is subtracted from each !!!

Replies: 506   Last Post: Nov 20, 2012 9:21 PM

Advanced Search

Back to Topic List Back to Topic List Jump to Tree View Jump to Tree View   Messages: [ Previous | Next ]
Ray Koopman

Posts: 3,383
Registered: 12/7/04
Re: Your assessment of current method of computing "u" relative to a
given group, e.g. S63

Posted: Jun 2, 2012 1:53 PM
  Click to see the message monospaced in plain text Plain Text   Click to reply to this topic Reply

On Jun 2, 8:30 am, djh <halitsk...@att.net> wrote:
> I assume from your last question that you're now either able or close
> to able to assess the current method of computing "u" relative to a
> given group (e.g. S63). By "current method", I mean this one (taken
> from a previous post):
>
> ******
> Suppose your boss gives you an encoded announcement in which:
>
> a) a 2-word sequence for AP appears just once and this 2-word
> sequence
> is att ccc (one of the above 63 2-word sequences)
>
> b) a 2-word sequence for AL appears just once and this 2-word
> sequence
> is gcg ctg (one of the above 63 2-word sequences)
>
> c) no other 2-word sequence from the above 63 occurs in the
> announcement (equivalently, you will typeset an announcement in which
> none of the above 49 2-letter sequences appears other than one AP and
> one AL.)
>
> Then the degree ?u? of average over-representation of the 63 2-word
> sequences in the announcement is computed by my PERL code as:
>
> 1+1 = 2 (1 of the 63 for AL and 1 of the 63 for AP)
>
> 1/12 + 1/8 = 5/24 (expected frequency for AL and expected frequency
> for AP)
>
> 2/(5/24) = 48/5 (actual freq / expected freq)
>
> (48/5)/2 = 48/10 = 4.8 = u (because only two dipeptide positions in
> the announcement were occupied by dipeptides in the 49 encoded by
> S63).
>
> *******
>
> If anything is unclear in the above, please let me know. And of
> course, if you think there's a more correct or better way to compute
> "u", I assume you'll let me know what method you would prefer. So
> long as the method gives us a measure u of over-representation
> relative to a given group (e.g. any one of S63, C711, S60, C537, S119,
> C1058), I will use whatever method you specify.
>
> Thanks again, Ray.
>
> ******


I'm getting there, but there are some intermediate steps I need to
take first to make sure I really understand things.
__________

If you still have the original 20K real messages that S63 is based
on, and if you should find yourself with nothing else to do, you
might try the following, which occurred to me last night while I
was thinking about the fact that S63 involved three arbitrary values
-- the inner and outer cutoffs, both of which were 95%, and the
significance level, which wasn't mentioned but was presumably 5% --
and wondering if there was a way to avoid them.

Make a 61 x 61 table whose entries are frequency counts of the all
the dicodons in the 20K messages. In each dicodon, the left codon
picks the row, and the right codon picks the column. (Side question:
what's the standard way to refer to the left and right codons in a
dicodon?)

Let fjk denote the frequency count in row j, column k.
Let rj denote the total in row j.
Let ck denote the total in col k.
Let n denote the grand total.
Make a new 61 x 61 table whose entries are

gjk = (n*fjk - rj*ck) / sqrt[ rj*(n-rj)*ck*(n-ck) ].

Each g is a correlation, so -1 <= g <= 1. (For significance testing,
refer g*sqrt[n-1] to the standard normal distribution.) g is a
measure of over/under-representation (+/-). Sort the g's, keeping
track of which dicodons they correspond to. Look at the distribution.
Is there any kind of break between the largest and the not-so-large,
that would let us "carve nature at its joints" if we santed to pick
the most over-represented dicodons, or would an arbitrary cutoff
have to be used? Where do the S63 dicodons come in the distribution?


Date Subject Author
5/14/12
Read The same four proportional weighting factors work for each
00/01/10/11 when 0.25 is subtracted from each !!!
Halitsky
5/14/12
Read Re: The same four proportional weighting factors work for each
00/01/10/11 when 0.25 is subtracted from each !!!
Ray Koopman
5/15/12
Read Re: The same four proportional weighting factors work for each
00/01/10/11 when 0.25 is subtracted from each !!!
Halitsky
5/15/12
Read Re: The same four proportional weighting factors work for each
00/01/10/11 when 0.25 is subtracted from each !!!
Ray Koopman
5/15/12
Read Re: The same four proportional weighting factors work for each
00/01/10/11 when 0.25 is subtracted from each !!!
Halitsky
5/15/12
Read Re: The same four proportional weighting factors work for each
00/01/10/11 when 0.25 is subtracted from each !!!
Ray Koopman
5/15/12
Read Re: The same four proportional weighting factors work for each
00/01/10/11 when 0.25 is subtracted from each !!!
Halitsky
5/16/12
Read Re: The same four proportional weighting factors work for each
00/01/10/11 when 0.25 is subtracted from each !!!
Ray Koopman
5/16/12
Read Is now also the time to try and equalize study and control group
sample sizes ??
Halitsky
5/16/12
Read Can you crunch this table to check for any issue which might affect
our "c/L" ratio?
Halitsky
5/18/12
Read Re: Can you crunch this table to check for any issue which might
affect our "c/L" ratio?
Ray Koopman
5/18/12
Read Re: Can you crunch this table to check for any issue which might
affect our "c/L" ratio?
Halitsky
5/18/12
Read Re: Can you crunch this table to check for any issue which might
affect our "c/L" ratio?
Ray Koopman
5/18/12
Read Re: Can you crunch this table to check for any issue which might
affect our "c/L" ratio?
Halitsky
5/19/12
Read Should I obtain "R" to do Bartlett's (or Fligner's) homoscedasticity
tests on the driver correlations?
Halitsky
5/19/12
Read Update re homoscedasticity testing: my wife can do them in MiniTab ....
Halitsky
5/20/12
Read 1) Sample "jackknife" table 2) Requisite n's 3) linear regression
comparisons; 4) sample equalization
Halitsky
5/20/12
Read Re: 1) Sample "jackknife" table 2) Requisite n's 3) linear regression
comparisons; 4) sample equalization
Ray Koopman
5/20/12
Read Re: 1) Sample "jackknife" table 2) Requisite n's 3) linear regression
comparisons; 4) sample equalization
Halitsky
5/21/12
Read Re: 1) Sample "jackknife" table 2) Requisite n's 3) linear regression
comparisons; 4) sample equalization
Ray Koopman
5/21/12
Read Re: 1) Sample "jackknife" table 2) Requisite n's 3) linear regression
comparisons; 4) sample equalization
Halitsky
5/21/12
Read Re: 1) Sample "jackknife" table 2) Requisite n's 3) linear regression
comparisons; 4) sample equalization
Ray Koopman
5/21/12
Read Re: 1) Sample "jackknife" table 2) Requisite n's 3) linear regression
comparisons; 4) sample equalization
Halitsky
5/22/12
Read Re: 1) Sample "jackknife" table 2) Requisite n's 3) linear regression
comparisons; 4) sample equalization
Ray Koopman
5/22/12
Read Re: 1) Sample "jackknife" table 2) Requisite n's 3) linear regression
comparisons; 4) sample equalization
Halitsky
5/24/12
Read Re: 1) Sample "jackknife" table 2) Requisite n's 3) linear regression
comparisons; 4) sample equalization
Ray Koopman
5/24/12
Read Re: 1) Sample "jackknife" table 2) Requisite n's 3) linear regression
comparisons; 4) sample equalization
Halitsky
5/24/12
Read Re: 1) Sample "jackknife" table 2) Requisite n's 3) linear regression
comparisons; 4) sample equalization
Halitsky
5/22/12
Read Re: 1) Sample "jackknife" table 2) Requisite n's 3) linear regression
comparisons; 4) sample equalization
Ray Koopman
5/24/12
Read New queston: need your advice on some "study group vs study group"
t-test results
Halitsky
5/25/12
Read Re: New queston: need your advice on some "study group vs study
group" t-test results
Ray Koopman
5/25/12
Read Re: New queston: need your advice on some "study group vs study
group" t-test results
Halitsky
5/25/12
Read Re: New queston: need your advice on some "study group vs study
group" t-test results
Ray Koopman
5/26/12
Read Re: New queston: need your advice on some "study group vs study
group" t-test results
Halitsky
5/28/12
Read Re: New queston: need your advice on some "study group vs study
group" t-test results
Ray Koopman
5/28/12
Read Re: New queston: need your advice on some "study group vs study
group" t-test results
Halitsky
5/28/12
Read Re: New queston: need your advice on some "study group vs study
group" t-test results
Halitsky
5/28/12
Read Re: New queston: need your advice on some "study group vs study
group" t-test results
Ray Koopman
5/28/12
Read Re: New queston: need your advice on some "study group vs study
group" t-test results
Halitsky
5/28/12
Read Re: New queston: need your advice on some "study group vs study
group" t-test results
Ray Koopman
5/29/12
Read Re: New queston: need your advice on some "study group vs study
group" t-test results
Ray Koopman
5/29/12
Read Re: New queston: need your advice on some "study group vs study
group" t-test results
Halitsky
5/29/12
Read Re: New queston: need your advice on some "study group vs study
group" t-test results
Halitsky
5/30/12
Read Re: New queston: need your advice on some "study group vs study
group" t-test results
Ray Koopman
5/30/12
Read The other matters you raised in your post last night
Halitsky
5/30/12
Read Oh! - you didn't mean a MANOVA - you meant a multiple linear
regression of one DV on 2 IV's. right?
Halitsky
5/30/12
Read Re: Oh! - you didn't mean a MANOVA - you meant a multiple linear
regression of one DV on 2 IV's. right?
Ray Koopman
5/29/12
Read New question re behavior of u and x2 driver when data selected for u
> 0 vs u > 1
Halitsky
5/30/12
Read Re: New question re behavior of u and x2 driver when data selected
for u > 0 vs u > 1
Ray Koopman
5/30/12
Read kclas/mols is the unit of "e", not u
Halitsky
5/30/12
Read Re: kclas/mols is the unit of "e", not u
Ray Koopman
5/30/12
Read Comparison of avg ln(c/u) on ln(c/L) coefficients to average "UCP's"
(u-computation probabilities)
Halitsky
5/30/12
Read Re: Comparison of avg ln(c/u) on ln(c/L) coefficients to average
"UCP's" (u-computation probabilities)
Ray Koopman
5/30/12
Read The relationship of UCP's to "u"
Halitsky
5/30/12
Read Naturally, a stupid minor error in my above post re UCP and u
Halitsky
5/30/12
Read Intuitive vs actual algorithm for computing u
Halitsky
5/31/12
Read Re: The relationship of UCP's to "u"
Ray Koopman
5/31/12
Read How to understand L, c, and u in terms of our allegory
Halitsky
5/31/12
Read Re: The relationship of UCP's to "u"
Ray Koopman
5/31/12
Read Re: The relationship of UCP's to "u"
Halitsky
5/31/12
Read Re: The relationship of UCP's to "u"
Halitsky
6/1/12
Read Re: The relationship of UCP's to "u"
Ray Koopman
6/1/12
Read Where the "63" came from
Halitsky
6/1/12
Read Re: Where the "63" came from
Ray Koopman
6/1/12
Read Re: Where the "63" came from
Halitsky
6/1/12
Read Re: Where the "63" came from
Halitsky
6/2/12
Read Re: The relationship of UCP's to "u"
Ray Koopman
6/2/12
Read Re: The relationship of UCP's to "u"
Halitsky
6/2/12
Read Your assessment of current method of computing "u" relative to a
given group, e.g. S63
Halitsky
6/2/12
Read Re: Your assessment of current method of computing "u" relative to a
given group, e.g. S63
Ray Koopman
6/2/12
Read Re: Your assessment of current method of computing "u" relative to a
given group, e.g. S63
Halitsky
6/2/12
Read Correction to presentation of "u" computation method
Halitsky
6/3/12
Read Re: Correction to presentation of "u" computation method
Ray Koopman
6/4/12
Read Re: Correction to presentation of "u" computation method
Ray Koopman
6/4/12
Read Re: Correction to presentation of "u" computation method
Halitsky
6/4/12
Read Clarification of the term "combined probability" in assertion (I) in
previous message
Halitsky
6/4/12
Read Please hold the fort ... I think I now see what you're saying re "u"
Halitsky
6/4/12
Read OK - here is exactly what my code is doing ... please evaluate
whether it correctly computes "u"
Halitsky
6/4/12
Read Re: OK - here is exactly what my code is doing ... please evaluate
whether it correctly computes "u"
Ray Koopman
6/5/12
Read Re: OK - here is exactly what my code is doing ... please evaluate
whether it correctly computes "u"
Ray Koopman
6/5/12
Read Re: OK - here is exactly what my code is doing ... please evaluate
whether it correctly computes "u"
Halitsky
6/6/12
Read Re: OK - here is exactly what my code is doing ... please evaluate
whether it correctly computes "u"
Ray Koopman
6/6/12
Read Re: OK - here is exactly what my code is doing ... please evaluate
whether it correctly computes "u"
Ray Koopman
6/6/12
Read Re: OK - here is exactly what my code is doing ... please evaluate
whether it correctly computes "u"
Halitsky
6/6/12
Read But I guess the u=0 problem would go away if we used u/c, e/c, and
L/c instead ....
Halitsky
6/6/12
Read Also, "e" has no meaning for segments in which u = 0 because N-yes = 0.
Halitsky
6/8/12
Read Correction of an egregiously bad "simplifying" assumption regarding
"c" ...
Halitsky
6/8/12
Read Re: Correction of an egregiously bad "simplifying" assumption
regarding "c" ...
Halitsky
6/9/12
Read Re: Correction of an egregiously bad "simplifying" assumption
regarding "c" ...
Ray Koopman
6/9/12
Read It's "rubber meets the road" time, Ray .... are these two regressions
sufficiently "different" ?
Halitsky
6/10/12
Read Restatement (improved) of previous correlation results after two
dicodon/UCP table errors corrected
Halitsky
6/11/12
Read Re: Restatement (improved) of previous correlation results after two
dicodon/UCP table errors corrected
Ray Koopman
6/11/12
Read Re: Restatement (improved) of previous correlation results after two
dicodon/UCP table errors corrected
Halitsky
6/11/12
Read Re: Restatement (improved) of previous correlation results after two
dicodon/UCP table errors corrected
Ray Koopman
6/11/12
Read Re: Restatement (improved) of previous correlation results after two
dicodon/UCP table errors corrected
Halitsky
6/11/12
Read Re: Restatement (improved) of previous correlation results after two
dicodon/UCP table errors corrected
Ray Koopman
6/11/12
Read 1) thanks for the additional summary stats; 2) next critical question
requiring your evaluation
Halitsky
6/12/12
Read Re: 1) thanks for the additional summary stats; 2) next critical
question requiring your evaluation
Ray Koopman
6/12/12
Read Re: 1) thanks for the additional summary stats; 2) next critical
question requiring your evaluation
Halitsky
6/13/12
Read Re: 1) thanks for the additional summary stats; 2) next critical
question requiring your evaluation
Ray Koopman
6/13/12
Read Re: 1) thanks for the additional summary stats; 2) next critical
question requiring your evaluation
Halitsky
6/14/12
Read New question: could u on c*L (no logs) be logarithmic (over ALL
length intervals)
gimpeltf@hotmail.com
6/15/12
Read Re: New question: could u on c*L (no logs) be logarithmic (over ALL
length intervals)
Ray Koopman
6/15/12
Read Re: New question: could u on c*L (no logs) be logarithmic (over ALL
length intervals)
Halitsky
6/15/12
Read Re: New question: could u on c*L (no logs) be logarithmic (over ALL
length intervals)
gimpeltf@hotmail.com
6/15/12
Read Re: New question: could u on c*L (no logs) be logarithmic (over ALL
length intervals)
Halitsky
6/15/12
Read Critical question re proper way to argue from slope CI
overlaps/non-overlaps among S63, S63R, and C711
Halitsky
6/16/12
Read Re: Critical question re proper way to argue from slope CI
overlaps/non-overlaps among S63, S63R, and C711
Ray Koopman
6/16/12
Read How stupid am I? Of course I had the Y's and the N's reversed ...
Halitsky
6/16/12
Read Re: How stupid am I? Of course I had the Y's and the N's reversed ...
Ray Koopman
6/16/12
Read Re: How stupid am I? Of course I had the Y's and the N's reversed ...
Halitsky
6/16/12
Read Correction: there is a slope CI overlap for a1 len 25 ln(c/u) on
ln(c/L) for S63 vs S63R
Halitsky
6/18/12
Read Current Status and New Question re ln(c/u) on ln(c/L)
Halitsky
6/18/12
Read Re: Current Status and New Question re ln(c/u) on ln(c/L)
Halitsky
6/18/12
Read Re: Current Status and New Question re ln(c/u) on ln(c/L)
Halitsky
6/19/12
Read All a1 data now complete - will be sending you summary and detail
file later today
Halitsky
6/19/12
Read Re: All a1 data now complete - will be sending you summary and detail
file later today
Halitsky
6/19/12
Read Re: All a1 data now complete - will be sending you summary and detail
file later today
Halitsky
6/22/12
Read Re: All a1 data now complete - will be sending you summary and detail
file later today
Ray Koopman
6/22/12
Read Re: All a1 data now complete - will be sending you summary and detail
file later today
Ray Koopman
6/22/12
Read Re: All a1 data now complete - will be sending you summary and detail
file later today
Ray Koopman
6/22/12
Read Re: All a1 data now complete - will be sending you summary and detail
file later today
Halitsky
6/23/12
Read Re: All a1 data now complete - will be sending you summary and detail
file later today
Ray Koopman
6/23/12
Read Re: All a1 data now complete - will be sending you summary and detail
file later today
Halitsky
6/24/12
Read Re: All a1 data now complete - will be sending you summary and detail
file later today
Ray Koopman
6/24/12
Read Re: All a1 data now complete - will be sending you summary and detail
file later today
Halitsky
6/24/12
Read Re: All a1 data now complete - will be sending you summary and detail
file later today
Ray Koopman
6/24/12
Read Re: All a1 data now complete - will be sending you summary and detail
file later today
Halitsky
6/25/12
Read Re: All a1 data now complete - will be sending you summary and detail
file later today
Ray Koopman
6/25/12
Read Re: All a1 data now complete - will be sending you summary and detail
file later today
Halitsky
6/25/12
Read Re: All a1 data now complete - will be sending you summary and detail
file later today
Halitsky
6/26/12
Read Re: All a1 data now complete - will be sending you summary and detail
file later today
Ray Koopman
6/26/12
Read Re: All a1 data now complete - will be sending you summary and detail
file later today
Halitsky
6/26/12
Read I see what you're saying now about the length intervals - it's MY
code that has shifted the intervals down !
Halitsky
6/26/12
Read Since I'm rerunning a1 and a3, do you have any thoughts on how I
should run ln(c/u) on ln(c/L) ?
Halitsky
6/26/12
Read I'm taking a day or two to automate some error-prone manual
processing steps
Halitsky
6/27/12
Read Would it be possible for me to do the "set-wise" regression
significance comparisons on my side?
Halitsky
6/27/12
Read Re: Would it be possible for me to do the "set-wise" regression
significance comparisons on my side?
Ray Koopman
6/27/12
Read Re: Would it be possible for me to do the "set-wise" regression
significance comparisons on my side?
Halitsky
6/27/12
Read One dumb question (just this one I hope) about getting p's from t's ...
Halitsky
6/27/12
Read Regarding previous post, would Excel TDIST function still be OK to
use here?
Halitsky
6/28/12
Read Re: Regarding previous post, would Excel TDIST function still be OK
to use here?
Ray Koopman
6/28/12
Read Re: Regarding previous post, would Excel TDIST function still be OK to use here?
Gaj Vidmar
6/28/12
Read Re: Regarding previous post, would Excel TDIST function still be OK
to use here?
Halitsky
6/28/12
Read Re: Regarding previous post, would Excel TDIST function still be OK
to use here?
Halitsky
6/28/12
Read Re: Regarding previous post, would Excel TDIST function still be
OK to use here?
Art Kendall
6/28/12
Read Re: Regarding previous post, would Excel TDIST function still be OK
to use here?
Halitsky
6/28/12
Read Re: Regarding previous post, would Excel TDIST function still be
OK to use here?
Art Kendall
6/28/12
Read Re: Regarding previous post, would Excel TDIST function still be OK
to use here?
Halitsky
6/28/12
Read Permissible comparative sizes of N's for sets being compared by your
t-test (and correction) algorithm
Halitsky
6/28/12
Read Re: Permissible comparative sizes of N's for sets being compared by
your t-test (and correction) algorithm
Ray Koopman
6/28/12
Read Next question: df's at each "layer" in a "Russian egg" significance
testing protocol
Halitsky
6/29/12
Read Re: Next question: df's at each "layer" in a "Russian egg"
significance testing protocol
Ray Koopman
6/29/12
Read Re: Next question: df's at each "layer" in a "Russian egg"
significance testing protocol
Halitsky
6/29/12
Read Typo in previous post - test-count = 24, not 48
Halitsky
6/30/12
Read Re: Next question: df's at each "layer" in a "Russian egg"
significance testing protocol
Ray Koopman
6/30/12
Read Re: Next question: df's at each "layer" in a "Russian egg"
significance testing protocol
Halitsky
6/30/12
Read But we COULD try t-testing corresponding sets of residuals per len
interval, if you think it valid to do so
Halitsky
7/1/12
Read Could you double-check me on the "custom t-test" calculation just
this once before you leave Tue?
Halitsky
7/2/12
Read Re: Could you double-check me on the "custom t-test" calculation just
this once before you leave Tue?
Ray Koopman
7/2/12
Read Re: Could you double-check me on the "custom t-test" calculation just
this once before you leave Tue?
Halitsky
7/2/12
Read I'm now a "Bonferroni-believer"!!! (At least till the results go the
WRONG way ...)
Halitsky
7/2/12
Read Re: I'm now a "Bonferroni-believer"!!! (At least till the results go
the WRONG way ...)
Halitsky
7/2/12
Read Bonferroni calculation results for ln(c/e) on ln(c/L) for Family 1
Halitsky
7/2/12
Read Re: I'm now a "Bonferroni-believer"!!! (At least till the results go
the WRONG way ...)
Ray Koopman
7/2/12
Read Corrected Bonferroni data for ln(ce) on ln(c/L) for Family 1, and 1
question on "random outliers"
Halitsky
7/4/12
Read Full results of test-suite on fold a1 AFTER intra-family Bonferroni correction
gimpeltf@hotmail.com
7/11/12
Read Post of offline email (7/11) re a1/a3/b1 results, for later
referenceability if needed
Halitsky
7/11/12
Read Re: Post of offline email (7/11) re a1/a3/b1 results, for later
referenceability if needed
Halitsky
7/17/12
Read Results of Bonferroni-corrected significance testing across all six
folds now complete
Halitsky
7/19/12
Read Quick specific question about bringing variance of a set of
observations to an expected value.
Halitsky
7/19/12
Read Re: Quick specific question about bringing variance of a set of
observations to an expected value.
Ray Koopman
7/19/12
Read Re: Quick specific question about bringing variance of a set of
observations to an expected value.
Halitsky
7/23/12
Read Can your custom t-test be run on means and variances as well as on
slopes and SE's or intercepts and SE's ?
Halitsky
7/23/12
Read Same question as last post, but with variance and sd of residuals
instead of mean and var ....
Halitsky
7/23/12
Read Example of variance and sd of residuals decreasing across length
intervals 1 to 12
Halitsky
7/23/12
Read Re: Example of variance and sd of residuals decreasing across length
intervals 1 to 12
Halitsky
7/23/12
Read Re: Can your custom t-test be run on means and variances as well as
on slopes and SE's or intercepts and SE's ?
Ray Koopman
7/23/12
Read OK - then how about this alternative approach to patterns of variances?
Halitsky
7/24/12
Read Re: OK - then how about this alternative approach to patterns of variances?
Ray Koopman
7/24/12
Read Re: OK - then how about this alternative approach to patterns of variances?
Ray Koopman
7/24/12
Read Thanks! but a quick question about thee construct "adjusted SD" ...
Halitsky
7/24/12
Read Oh wait a second ... by "adjusted SD" do you simply mean the one with
N-1 rather than N in the denominator?
Halitsky
7/24/12
Read Re: Oh wait a second ... by "adjusted SD" do you simply mean the one
with N-1 rather than N in the denominator?
Ray Koopman
7/24/12
Read So N-2 instead of N-1 ???? (sorry to be a nuisance .)
Halitsky
7/24/12
Read Re: So N-2 instead of N-1 ???? (sorry to be a nuisance .)
Ray Koopman
7/24/12
Read Re: So N-2 instead of N-1 ???? (sorry to be a nuisance .)
Halitsky
7/24/12
Read is there a way you "adjudicate" a case like this (0.00002 slope CI
overlap and 0.00017 intcpt CI overlap)?
Halitsky
7/24/12
Read Re: is there a way you "adjudicate" a case like this (0.00002 slope
CI overlap and 0.00017 intcpt CI overlap)?
Ray Koopman
7/24/12
Read Re: is there a way you "adjudicate" a case like this (0.00002 slope
CI overlap and 0.00017 intcpt CI overlap)?
Halitsky
7/24/12
Read Re: is there a way you "adjudicate" a case like this (0.00002 slope
CI overlap and 0.00017 intcpt CI overlap)?
Ray Koopman
7/25/12
Read Results of fold-discrimination analysis via adjusted SD analysis
Halitsky
7/26/12
Read A follow-up note about fold c1 in fold-discrimination matrix
presented in last post
Halitsky
7/26/12
Read Re: A follow-up note about fold c1 in fold-discrimination matrix
presented in last post
Ray Koopman
7/26/12
Read Re: A follow-up note about fold c1 in fold-discrimination matrix
presented in last post
Halitsky
7/27/12
Read Re: A follow-up note about fold c1 in fold-discrimination matrix
presented in last post
Ray Koopman
7/27/12
Read Re: A follow-up note about fold c1 in fold-discrimination matrix
presented in last post
Halitsky
7/27/12
Read Results of “Cross-Filtering” Approach Suggested
in Previous Post
Halitsky
7/27/12
Read Test post to check line size - please ignore
Halitsky
8/1/12
Read If I'd only thought more carefully about a suggestion you made 2
months ago ....
Halitsky
8/2/12
Read Two really quick questions about Bonferroni corrections
Halitsky
8/2/12
Read Re: Two really quick questions about Bonferroni corrections
Ray Koopman
8/2/12
Read Re: Two really quick questions about Bonferroni corrections
Halitsky
8/2/12
Read Re: Two really quick questions about Bonferroni corrections
Richard Ulrich
8/3/12
Read Re: Two really quick questions about Bonferroni corrections
Halitsky
8/3/12
Read QUICK question about whether a certain result is/isn
’t sample-size independent
Halitsky
8/5/12
Read I'm proceeding without your answer to my last question, BUT ...
Halitsky
8/5/12
Read Re: I'm proceeding without your answer to my last question, BUT ...
Ray Koopman
8/5/12
Read Re: I'm proceeding without your answer to my last question, BUT ...
Halitsky
8/5/12
Read Example tabulation of N's and QUICK question re a possible sampling
protocol ...
Halitsky
8/5/12
Read PS - I think you suggested a protocol like the above some time ago ...
Halitsky
8/7/12
Read Example of Bonferroni ranking (presumably) independent of sample size
Halitsky
8/15/12
Read Re: Example tabulation of N's and QUICK question re a possible
sampling protocol ...
Ray Koopman
8/15/12
Read Before I go to bootstrapping, could you evaluate one last simple proposal?
Halitsky
8/15/12
Read Re: Before I go to bootstrapping, could you evaluate one last simple proposal?
Ray Koopman
8/15/12
Read OOPS! But is there another way to test the two sets of eigenvalues ?????
Halitsky
8/16/12
Read Re: OOPS! But is there another way to test the two sets of
eigenvalues ?????
Ray Koopman
8/16/12
Read OK - it looks like we can wrap this thread up ...
Halitsky
8/8/12
Read Re: QUICK question about whether a certain result is
/isn’t sample-size independent
Ray Koopman
8/8/12
Read When you “return”, please note the LATEST propos
ed design in my posts of 8/5@720&738 and 8/7@731
Halitsky
8/15/12
Read Re: When you “return”, please note the LATEST pr
oposed design in my posts of 8/5@720&738 and 8/7@731
Ray Koopman
8/15/12
Read My question was whether the intercept of x1 on x2,x3 is common to
BOTH regression lines of this multiple regression.
Halitsky
8/15/12
Read Re: My question was whether the intercept of x1 on x2,x3 is common to
BOTH regression lines of this multiple regression.
Ray Koopman
8/15/12
Read My problem is understanding the single intercept that Excel reports
for x1 on (x2,x3)
Halitsky
8/15/12
Read Re: My problem is understanding the single intercept that Excel
reports for x1 on (x2,x3)
Ray Koopman
8/15/12
Read Re: My problem is understanding the single intercept that Excel
reports for x1 on (x2,x3)
Halitsky
8/13/12
Read Re: QUICK question about whether a certain result is
/isn’t sample-size independent
Ray Koopman
8/13/12
Read Re: QUICK question about whether a certain result is
/isn’t sample-size independent
Halitsky
8/22/12
Read Would you have time to devise the test for comparing a simple and a
multiple regression?
Halitsky
8/23/12
Read Re: Would you have time to devise the test for comparing a simple and
a multiple regression?
Ray Koopman
8/23/12
Read We may be talking at cross-purposes - I'm not sure ...
Halitsky
8/24/12
Read Re: We may be talking at cross-purposes - I'm not sure ...
Ray Koopman
8/25/12
Read I trust your ability to see statistical clarity through the messiness
Halitsky
8/31/12
Read Re: I trust your ability to see statistical clarity through the messiness
Ray Koopman
8/31/12
Read Thanks for characterizing the nature of the d.v to look for ...
Halitsky
8/31/12
Read Correction of two typo's in prior post
Halitsky
9/3/12
Read Re: Thanks for characterizing the nature of the d.v to look for ...
Ray Koopman
9/3/12
Read If the results herein are technically legit, then you're easy a
genius, or very very experienced, or both.
Halitsky
9/4/12
Read Re: If the results herein are technically legit, then you're easy a
genius, or very very experienced, or both.
Ray Koopman
9/4/12
Read Re: If the results herein are technically legit, then you're easy a
genius, or very very experienced, or both.
Halitsky
9/5/12
Read Re: If the results herein are technically legit, then you're easy a
genius, or very very experienced, or both.
Ray Koopman
9/5/12
Read Sorry - I misunderstood your term "error variances" in your post of 8/24@823.
Halitsky
9/5/12
Read A technical problem with Anderson's LineFit module is what confused
me about the "variances"
Halitsky
9/5/12
Read Can N-related loss of confidence be quantified for each result of
your original custom heteroscedastic T?
Halitsky
9/6/12
Read Re: Can N-related loss of confidence be quantified for each result of
your original custom heteroscedastic T?
Ray Koopman
9/6/12
Read Re: Can N-related loss of confidence be quantified for each result of
your original custom heteroscedastic T?
Halitsky
9/6/12
Read Two more follow-on questions to the two I just asked in my last post
of 9/6@5:36 am
Halitsky
9/6/12
Read PS - Our control set expansion from 1 to 3 is not only scientifically
justifiable, but scientifically obligatory ....
gimpeltf@hotmail.com
9/6/12
Read Re: PS - Our control set expansion from 1 to 3 is not only
scientifically justifiable, but scientifically obligatory ....
Ray Koopman
9/6/12
Read Sorry for the obscure presentation of the sets - here's a clear "laydown"
Halitsky
9/9/12
Read New example summary and detail files
Halitsky
9/9/12
Read Re: New example summary and detail files
Ray Koopman
9/9/12
Read Re: New example summary and detail files
Halitsky
9/10/12
Read Re: New example summary and detail files
Ray Koopman
9/10/12
Read Thanks... three questions re your last post (quickly answerable, I
hope ...)
Halitsky
9/11/12
Read Re: Thanks... three questions re your last post (quickly answerable,
I hope ...)
Ray Koopman
9/11/12
Read Thanks (correction of variance and CIs of slope differences)
Halitsky
9/11/12
Read Could you check (and explain) this “negative t”
case?
Halitsky
9/11/12
Read Re: Could you check (and explain) this “negative t
” case?
Ray Koopman
9/11/12
Read Re: Could you check (and explain) this “negative t
” case?
Halitsky
9/9/12
Read Custom heteroscedastic test is working in Perl but with integer "ft"
and 6 decimal places for returned p;
Halitsky
9/11/12
Read Re: Custom heteroscedastic test is working in Perl but with integer
"ft" and 6 decimal places for returned p;
Ray Koopman
9/11/12
Read Yes - the opensource lib of Perl math modules does have a betainc ...
Halitsky
9/11/12
Read Re: Yes - the opensource lib of Perl math modules does have a betainc ...
Ray Koopman
9/11/12
Read Re: Yes - the opensource lib of Perl math modules does have a betainc ...
Halitsky
9/12/12
Read If I've done the CI's correctly, then initial results are encouraging
(I think)
Halitsky
9/12/12
Read Re: If I've done the CI's correctly, then initial results are
encouraging (I think)
Ray Koopman
9/12/12
Read Re: If I've done the CI's correctly, then initial results are
encouraging (I think)
Halitsky
9/13/12
Read Re: If I've done the CI's correctly, then initial results are
encouraging (I think)
Ray Koopman
9/13/12
Read But 1.6457 is NOT the value you want , correct ? (Because Excel
T.INV.T returns your 1.96127986 )
Halitsky
9/13/12
Read Re: But 1.6457 is NOT the value you want , correct ? (Because Excel
T.INV.T returns your 1.96127986 )
Ray Koopman
9/13/12
Read Yes - you're correct! - Even for the "grown-up" inverse t, I have to
give it 1 minus HALF of (1 minus desired CI)
Halitsky
9/13/12
Read Re: If I've done the CI's correctly, then initial results are encouraging
(I think)
Bruce Weaver
9/13/12
Read Re: If I've done the CI's correctly, then initial results are
encouraging (I think)
Ray Koopman
9/13/12
Read Re: If I've done the CI's correctly, then initial results are encouraging
(I think)
Bruce Weaver
9/13/12
Read Which do YOU think are the "n independent parameters" for the
multiple CI evaluation?
Halitsky
9/14/12
Read Re: Which do YOU think are the "n independent parameters" for the
multiple CI evaluation?
Ray Koopman
9/14/12
Read OK. Then seven separate analyes with n = 12 for the ".95^(1/n)" CIs.
Halitsky
9/14/12
Read Re: OK. Then seven separate analyes with n = 12 for the ".95^(1/n)" CIs.
Ray Koopman
9/14/12
Read Re: OK. Then seven separate analyes with n = 12 for the ".95^(1/n)" CIs.
Halitsky
9/14/12
Read Am sending three files offline with n=1,12,73 for ".95^(1/n)"
Halitsky
9/14/12
Read Re: Am sending three files offline with n=1,12,73 for ".95^(1/n)"
Ray Koopman
9/14/12
Read Re: Am sending three files offline with n=1,12,73 for ".95^(1/n)"
Halitsky
9/15/12
Read Re: Am sending three files offline with n=1,12,73 for ".95^(1/n)"
Ray Koopman
9/15/12
Read Re: Am sending three files offline with n=1,12,73 for ".95^(1/n)"
Ray Koopman
9/15/12
Read Yes - your last instructions give me same tCI's as yours - sorry for
my confusion
Halitsky
9/15/12
Read Re: Am sending three files offline with n=1,12,73 for ".95^(1/n)"
Ray Koopman
9/15/12
Read k-2 for Re, Ru coeff's vs k-3 for Reu coeffs
Halitsky
9/15/12
Read k = 3 changes df from ~1804 to ~1785 for the "euSu" coefficient
Halitsky
9/15/12
Read Am sending a zip file with complete CI results for a1 fold across all
cells of the design ...
Halitsky
9/15/12
Read Had you already looked at the data or did you simply have a great
experience-based hunch?
Halitsky
9/15/12
Read CI Partition analysis for all seven co-efficients (a1 fold only)
Halitsky
9/17/12
Read I've got the CI data for the other 5 folds, but don't want to post
them if the "thirds" approach is wrong ...
Halitsky
9/20/12
Read Re: I've got the CI data for the other 5 folds, but don't want to
post them if the "thirds" approach is wrong ...
Ray Koopman
9/20/12
Read 1) Thanks for the algorithm! 2) your initial a1 plot should please
both JRF and AML
Halitsky
9/20/12
Read Since your plot was for the eI coefficient, my previous comments were
relative only to THAT coefficient
Halitsky
9/21/12
Read Here are the plots for all 7 coeff's for the a1 fold, along with some comments/questions
Halitsky
9/21/12
Read Re: Here are the plots for all 7 coeff's for the a1 fold, along with
some comments/questions
Ray Koopman
9/21/12
Read CI plots for a3 fold
Halitsky
9/21/12
Read CI plots for b1 fold
Halitsky
9/21/12
Read CI plots for b47 fold
Halitsky
9/21/12
Read CI plots for c1 fold
Halitsky
9/21/12
Read CI plots for c2 fold
Halitsky
9/22/12
Read Tabulation of S:C cell pairs which correlate with non-overlapping CI's
Halitsky
9/23/12
Read Re: Tabulation of S:C cell pairs which correlate with non-overlapping CI's
Ray Koopman
9/23/12
Read OK - then here are four specific questions involving patterns (not
counts) of significant results
Halitsky
9/23/12
Read Please discard all posted CI results!! My computation of LL and HH
was wrong !
Halitsky
9/23/12
Read Re: Please discard all posted CI results!! My computation of LL and
HH was wrong !
Ray Koopman
9/23/12
Read Thanks for that clarification: will recompute with 72 CI's per plot
and (LL,HH) per each 72
Halitsky
9/23/12
Read I've sent you the CI plots per coefficient, and am now ready to throw
in the towel
Halitsky
9/24/12
Read Re: I've sent you the CI plots per coefficient, and am now ready to
throw in the towel
Ray Koopman
9/24/12
Read Re: I've sent you the CI plots per coefficient, and am now ready to
throw in the towel
Halitsky
9/24/12
Read Please review the eS, eI, and euI plots that I've sent you offline
for the "uA" design (multiplicity = 36)
Halitsky
9/25/12
Read Re: Please review the eS, eI, and euI plots that I've sent you
offline for the "uA" design (multiplicity = 36)
Ray Koopman
9/25/12
Read Thanks for taking a look at the uA plots; please permit some
questions about the custom t-test itself
Halitsky
9/25/12
Read one more question (sorry!): can N be validly regressed on (e,u,L)?
Halitsky
9/27/12
Read Re: one more question (sorry!): can N be validly regressed on (e,u,L)?
Ray Koopman
9/27/12
Read Re: Thanks for taking a look at the uA plots; please permit some
questions about the custom t-test itself
Ray Koopman
9/27/12
Read OK, then please do NOT misinterpret my motivation for asking this
next question ....
Halitsky
9/28/12
Read Re: OK, then please do NOT misinterpret my motivation for asking this
next question ....
Ray Koopman
9/28/12
Read Re: OK, then please do NOT misinterpret my motivation for asking this
next question ....
Halitsky
9/29/12
Read When a small set of p's passes the IOTT, is a Bonferroni correction
still required?
Halitsky
9/30/12
Read Can the argument from means be tightened by t-testin
g means “across N, R” as well as “across S, C”?
Halitsky
9/30/12
Read Augmented tables for remaining five folds (results of t-testing
mean(ln(e)) ...
Halitsky
10/1/12
Read Re: Can the argument from means be tightened by t-te
sting means “across N, R” as well as “across S, C”?
Ray Koopman
10/1/12
Read Thanks! Also, please note I'm going to bite the bullet and
"Fisher-Yates" the real data 20 times ....
Halitsky
10/1/12
Read Minor typo in previous post (Astrings instead of Cstrings in one place)
Halitsky
10/1/12
Read Re: Thanks! Also, please note I'm going to bite the bullet and
"Fisher-Yates" the real data 20 times ....
Ray Koopman
10/1/12
Read Yes. please (more detail on the "build-up" of variances and df's)
Halitsky
10/2/12
Read Re: Yes. please (more detail on the "build-up" of variances and df's)
Ray Koopman
10/2/12
Read 1) my restatement of your explanation of the formula; 2)
clarification of "random vs non-random"
Halitsky
10/2/12
Read Re: 1) my restatement of your explanation of the formula; 2)
clarification of "random vs non-random"
Ray Koopman
10/2/12
Read Ahh! "4" meant "point 4", not "4 of something" ...
Halitsky
10/3/12
Read Could you check this one result before I run the 1728?
Halitsky
10/4/12
Read Also, please note I used raw e's, not ln(e)'s ...
Halitsky
10/4/12
Read Also: NOT length interval = 25. Length interval = 1 (lowest)
Halitsky
10/4/12
Read Re: Also: NOT length interval = 25. Length interval = 1 (lowest)
Ray Koopman
10/4/12
Read Unfortunately, I think we now have an absolute rule-out, at least
when trying to use mean of raw e ...
Halitsky
10/4/12
Read Forgot to ask: is it worth "leveraging-up" even though no "base" p
per length interval is < 0.23 ?
Halitsky
10/5/12
Read Re: Forgot to ask: is it worth "leveraging-up" even though no "base"
p per length interval is < 0.23 ?
Ray Koopman
10/5/12
Read Have taken logs; will now leverage; while I'm doing so, could you
consider this possible design error?
Halitsky
10/5/12
Read We MAY be on the right track; if so, how can I possibly thank/repay you?
Halitsky
10/5/12
Read Technical follow-up question re “n choose 2” ana
lysis of table in last post
Halitsky
10/5/12
Read Please forgive the "rose-colored glasses" typo of 0.00013 in my last
post; of course it's 0.0013
Halitsky
10/5/12
Read Re: Please forgive the "rose-colored glasses" typo of 0.00013 in my
last post; of course it's 0.0013
Ray Koopman
10/5/12
Read Re: Technical follow-up question re “n choose 2”
analysis of table in last post
Ray Koopman
10/5/12
Read I have to ask you to clarify your response, and also specify how to
"look at" the t numerator ...
Halitsky
10/5/12
Read Here is the table of t-numerators, sorted in DECREASING order within
fold ....
Halitsky
10/6/12
Read Correct table of "t-numerators" (prior table was t's themselves); but
0.0013 result still holds.
Halitsky
10/6/12
Read Here are the 10 values of the Ornstein-Fresco index used in our
calculation of raw e
Halitsky
10/6/12
Read Calculations in my last post were clearly wrong (re t-num diff's vs
Ornstein-Fresco index spread)
Halitsky
10/13/12
Read Re: Here are the 10 values of the Ornstein-Fresco index used in our
calculation of raw e
Ray Koopman
10/13/12
Read Re: Here are the 10 values of the Ornstein-Fresco index used in our
calculation of raw e
Halitsky
10/13/12
Read Yes - but actually, all the t's should be changed to u's because the
numbers are for RNA, not DNA
Halitsky
10/13/12
Read Re: Yes - but actually, all the t's should be changed to u's because
the numbers are for RNA, not DNA
Ray Koopman
10/13/12
Read Re: Yes - but actually, all the t's should be changed to u's because
the numbers are for RNA, not DNA
Halitsky
10/13/12
Read I looked at the Gibley Ahlquist passage - they're dealing with the
O-F indices on the DNA side of the house ...
Halitsky
10/14/12
Read 1) PDF's of the O-F papers have been emailed; 2) spoke with Jacques,
will speak further with him on Tue
Halitsky
10/14/12
Read Re: 1) PDF's of the O-F papers have been emailed; 2) spoke with
Jacques, will speak further with him on Tue
Ray Koopman
10/14/12
Read Correct p's and df's for Y, Z, Y-Z for 1:R1, 2:R2, 3:R3
Halitsky
10/15/12
Read Finally - a result consistent with our earliest logistic regression
alignability result using a predictor derived from eS of Re
Halitsky
10/15/12
Read Re: Finally - a result consistent with our earliest logistic
regression alignability result using a predictor derived from eS of Re
Ray Koopman
10/15/12
Read An attempt to appopriately organize and caption the three previous tables
Halitsky
10/16/12
Read Relationship of eS to ln(e) (via IOTT and Linear Regression)
Halitsky
10/16/12
Read Oops - sorry: one exception to the IOTT result for uH stated in
previous post
Halitsky
10/18/12
Read Re: An attempt to appopriately organize and caption the three
previous tables
Ray Koopman
10/18/12
Read You really are too kind ... seriously ...
Halitsky
10/18/12
Read Re: You really are too kind ... seriously ...
Ray Koopman
10/18/12
Read 1)Thanks! ( I was taking varY and VarZ to the .05 power, not the .5
power !);
Halitsky
10/20/12
Read Please consider this analysis of mean ln(e) sums and mean eS sums.
Halitsky
10/20/12
Read Your basic plot algorithm MAY help inject a significant dose of
empirical reality into the analysis
Halitsky
10/20/12
Read Please consider these interaction p’s in re our st
rategy for prediction of structural alignability via logisti
c regression.
Halitsky
10/21/12
Read Now that you’ve shown how to tell “where”, can
you show how to tell “why”?
Halitsky
10/18/12
Read Jacques' summary response re scaling of e
Halitsky
10/22/12
Read Re: Jacques' summary response re scaling of e
Ray Koopman
10/22/12
Read Do you have any suggestions about how to "convert" the O-F index for
our purposes?
Halitsky
10/22/12
Read Re: Do you have any suggestions about how to "convert" the O-F index
for our purposes?
Ray Koopman
10/22/12
Read Re: Do you have any suggestions about how to "convert" the O-F index
for our purposes?
Halitsky
10/23/12
Read Re: Do you have any suggestions about how to "convert" the O-F index
for our purposes?
Ray Koopman
10/23/12
Read Believe it or not, I can now answer your questions re the O-F index
clearly and concisely.
Halitsky
10/23/12
Read Re: Believe it or not, I can now answer your questions re the O-F
index clearly and concisely.
Ray Koopman
10/23/12
Read Whoops - sorry! I recapped my code from memory incorrectly.
Halitsky
10/24/12
Read Re: Whoops - sorry! I recapped my code from memory incorrectly.
Ray Koopman
10/24/12
Read Re: Whoops - sorry! I recapped my code from memory incorrectly.
Halitsky
10/24/12
Read Have heard from JRF; he's OK with a composite OFI per dicodon
(defined as a sum, not an average.)
Halitsky
10/25/12
Read Re: Have heard from JRF; he's OK with a composite OFI per dicodon
(defined as a sum, not an average.)
Ray Koopman
10/25/12
Read Re: Have heard from JRF; he's OK with a composite OFI per dicodon
(defined as a sum, not an average.)
Halitsky
10/25/12
Read Minor correction to last post regarding "e = S / 3"
Halitsky
10/25/12
Read Re: Have heard from JRF; he's OK with a composite OFI per dicodon
(defined as a sum, not an average.)
Ray Koopman
10/25/12
Read Re: Have heard from JRF; he's OK with a composite OFI per dicodon
(defined as a sum, not an average.)
Halitsky
10/25/12
Read You've probably realized by now that the dH for a single doublet ONLY
has scientific meaning as a unit in an additive calculation.
Halitsky
10/25/12
Read Re: Have heard from JRF; he's OK with a composite OFI per dicodon
(defined as a sum, not an average.)
Ray Koopman
10/25/12
Read You’re correct: the dhbarcalc for the equation IS
an AVERAGE, not a SUM.
Halitsky
10/26/12
Read Am sending off-line a "csv" file with the Tm's and Kelvin Tm's for
the 3721 "non-stop" dicodons
Halitsky
10/27/12
Read Re: Am sending off-line a "csv" file with the Tm's and Kelvin Tm's
for the 3721 "non-stop" dicodons
Ray Koopman
10/28/12
Read Interaction p’s using “new e”
Halitsky
10/28/12
Read Re: Interaction p’s using “new e”
Ray Koopman
10/28/12
Read Re: Interaction p’s using “new e”
Ray Koopman
10/28/12
Read I figured the Bonferroni-correction was coming ...
Halitsky
10/29/12
Read My naive opinion is that "new e" clarifies the situation considerably ...
Halitsky
10/29/12
Read Correction of two obvious typos in Ru section of last posted table
Halitsky
10/29/12
Read Re: My naive opinion is that "new e" clarifies the situation
considerably ...
Ray Koopman
10/29/12
Read Re: My naive opinion is that "new e" clarifies the situation
considerably ...
Halitsky
10/29/12
Read Re: My naive opinion is that "new e" clarifies the situation
considerably ...
Ray Koopman
10/29/12
Read Re: My naive opinion is that "new e" clarifies the situation
considerably ...
Halitsky
10/29/12
Read Is there a standard ranking function which takes into account |x-y|
as well as x/y ?
Halitsky
10/31/12
Read Re: Is there a standard ranking function which takes into account
|x-y| as well as x/y ?
Ray Koopman
10/31/12
Read Please agree to or improve on the following interaction nomenclature
gimpeltf@hotmail.com
11/2/12
Read Re: Please agree to or improve on the following interaction nomenclature
Ray Koopman
11/2/12
Read Now that I “understand” the effect/interaction s
caffold, I can better frame objectives/goals.
Halitsky
11/2/12
Read Minor but confusing typo in (V,B,4) of last post
Halitsky
11/4/12
Read I MAY have a correct generalization, but don’t wan
t to post prematurely.
Halitsky
11/5/12
Read Re: I MAY have a correct generalization, but don’t
want to post prematurely.
Ray Koopman
11/5/12
Read My apologies for the "overload", and thanks for the "go-ahead"
Halitsky
11/7/12
Read Follow-up question I forgot to ask re ln(L) on (ln(c),ln(e)) and
ln(L) on (ln(c),ln(u)) ...
Halitsky
11/8/12
Read Nope – those 2 new regressions don’t work out at
all
Halitsky
11/9/12
Read Within length intervals, it IS OK to consider ln(c) on ln(e) and
ln(c) on ln(u).
Halitsky
11/10/12
Read Re: My apologies for the "overload", and thanks for the "go-ahead"
Ray Koopman
11/10/12
Read Thanks for the review of the putative generalization ...
Halitsky
11/10/12
Read Please note that one column of the second table in my last post "wrapped"
Halitsky
11/11/12
Read Is this 1997 study a good practical example of a trivariate
distribution analysis?
Halitsky
11/11/12
Read Re: Is this 1997 study a good practical example of a trivariate
distribution analysis?
Ray Koopman
11/12/12
Read 1) thanks; 2) the simplified regressions in my prior post; 3) Ogle's
algorithm for generating bivariate normals with specific correlations
Halitsky
11/12/12
Read Re: 1) thanks; 2) the simplified regressions in my prior post; 3)
Ogle's algorithm for generating bivariate normals with specific correlations
Ray Koopman
11/12/12
Read Thansk for taking the time to try and decipher my reasoning re 3-ways
associated with the simplified model ...
Halitsky
11/12/12
Read Not sure if last post went thru, so am reposting - may result in a duplicate
Halitsky
11/16/12
Read Re: Not sure if last post went thru, so am reposting - may result in
a duplicate
Ray Koopman
11/16/12
Read Your latest assessment of the situation
Halitsky
11/16/12
Read Your "3-way" methodology has revealed two data tendencies wondrously
subtle and elegant ....
Halitsky
11/18/12
Read 1) Good news: two scientifically reasonable results;
2) Bad news: they don’t withstand Bonferroni correction
Halitsky
11/18/12
Read Have sent you off-line a file of uL and uH medians (
I’ve changed to your median approach)
Halitsky
11/16/12
Read A computationally cheap way to visualize “(e,c,u)|
L” (cheaper than trivariate analysis)
Halitsky
11/16/12
Read Re: A computationally cheap way to visualize “(e,c
,u)|L” (cheaper than trivariate analysis)
Ray Koopman
11/16/12
Read I had a sneaking suspicion you were going to insist on mapping to [0,1].
Halitsky
11/19/12
Read Re: I had a sneaking suspicion you were going to insist on mapping to [0,1].
Ray Koopman
11/19/12
Read Thanks for your formalization of the “centroid”
mechanics and the estiamte of their possible utility
Halitsky
11/20/12
Read Re: Thanks for your formalization of the “centroid
” mechanics and the estiamte of their possible utility
Ray Koopman
11/20/12
Read 1) Thanks for reviewing the “u-relativization” p
roposal; 2) question re “centroids”
Halitsky
11/20/12
Read Re: 1) Thanks for reviewing the “u-relativization
” proposal; 2) question re “centroids”
Ray Koopman
11/20/12
Read Responses (various) to your last (11/20:454pm)
Halitsky
10/23/12
Read A postscript from Jacques ...
Halitsky
10/8/12
Read If we had data for 20 folds instead of 6, could we argue from the new
result herein ?
Halitsky
10/8/12
Read The result in my previous post suggests a design sim
plification that MAY yield CI plots with decent “splits”
Halitsky
10/9/12
Read Re: If we had data for 20 folds instead of 6, could we argue from the
new result herein ?
Ray Koopman
10/9/12
Read Reply to your post of 10/9 at 3:54pm
Halitsky
10/10/12
Read Re: Reply to your post of 10/9 at 3:54pm
Ray Koopman
10/10/12
Read Please check if I understand how to get (Y,varY, dfY) at uL
Halitsky
10/11/12
Read Re: Please check if I understand how to get (Y,varY, dfY) at uL
Ray Koopman
10/11/12
Read Thanks(!) for the clarification – please just chec
k me now on the very first step.
Halitsky
10/11/12
Read Re: Thanks(!) for the clarification – please just
check me now on the very first step.
Ray Koopman
10/12/12
Read Re: Thanks(!) for the clarification – please just
check me now on the very first step.
Ray Koopman
10/12/12
Read Thanks again; now, please permit a question about applicability of
the protocol to coefficients
Halitsky
10/12/12
Read Re: Thanks again; now, please permit a question about applicability
of the protocol to coefficients
Ray Koopman
10/12/12
Read Are these "good enough" to keep going: pY at uL~.95, pZ at uH~.33,
p(Y-Z)~.23 ???
Halitsky
10/12/12
Read Forgot to mention: as you expected, df(Y), df(Z), and df(Y-Z) were
huge ...
Halitsky
10/12/12
Read Re: Forgot to mention: as you expected, df(Y), df(Z), and df(Y-Z)
were huge ...
Ray Koopman
10/12/12
Read I assume you're referring to the null hypothesis here, i.e. that the
means are the same ...
Halitsky
10/12/12
Read Re: Are these "good enough" to keep going: pY at uL~.95, pZ at
uH~.33, p(Y-Z)~.23 ???
Ray Koopman
10/12/12
Read Re vars vs SE's ...
Halitsky
10/10/12
Read Regarding the question of fold-generalizability, here's the question ...
Halitsky
10/11/12
Read Re: Regarding the question of fold-generalizability, here's the
question ...
Ray Koopman
10/9/12
Read S and new C subsets for dicodon set 1 (showing that
expected u’s are now the same for S and C)
Halitsky
10/1/12
Read Two follow-on POSSIBLE “$64K” questions (I’m H
OPING you MAY agree that they are!)
Halitsky
9/21/12
Read Before doing the other five folds, I should state "a priori" what
Jacques/Arthur would like to see ...
Halitsky
9/21/12
Read One other "a prior" note re Jacques' and Arthur's devoutest hope ...
Halitsky
8/25/12
Read For multiple linear regressions, are definitions same for SlopeVar,
IntcptVar, and (Sl,Int)Covar???
Halitsky
8/26/12
Read Re: For multiple linear regressions, are definitions same for
SlopeVar, IntcptVar, and (Sl,Int)Covar???
Ray Koopman
8/26/12
Read For our multiple regression ,will SE's of coefficients (including
intercept) suffice for now?
Halitsky
8/26/12
Read As you can see from my offline emails with Ivo Welch, I CAN get the
covariance matrix that's required ...
Halitsky
8/27/12
Read Re: As you can see from my offline emails with Ivo Welch, I CAN get
the covariance matrix that's required ...
Ray Koopman
8/27/12
Read Re: As you can see from my offline emails with Ivo Welch, I CAN get
the covariance matrix that's required ...
Halitsky
8/27/12
Read So far so good - Ivo's coefficients are same as yours for your test data
Halitsky
8/28/12
Read Quick question re multiplying AX'y to get b.
Halitsky
8/28/12
Read Thanks for the cautionary note, but here's why there's no cause for
concern ...
Halitsky
8/31/12
Read Would more pairs of files help you determine whether you can
construct the desired test?
Halitsky
8/29/12
Read List of Descriptors for Forthcoming "Structural Parameters" Results Files
Halitsky
8/28/12
Read I've got v and C = vA. and they match yours !!!!!
Halitsky
7/25/12
Read Test post only to check available horizontal space ... please ignore
Halitsky
6/27/12
Read Prelim results for ln(c/u) on ln(c/L) for S63 computed for u_low,
u_high, and u_all
Halitsky
6/27/12
Read Re: Prelim results for ln(c/u) on ln(c/L) for S63 computed for u_low,
u_high, and u_all
Halitsky
6/22/12
Read Re: All a1 data now complete - will be sending you summary and detail
file later today
Halitsky
6/22/12
Read Re: All a1 data now complete - will be sending you summary and detail
file later today
Halitsky
6/2/12
Read The expected frequency tables for S60, C537, S119, and C1058 are all
OK - I've double-checked them.
Halitsky
6/3/12
Read Re: The relationship of UCP's to "u"
Ray Koopman
5/30/12
Read Please ignore 4 tables in last post - here are the tables WITHOUT
three clerical errors.
Halitsky
5/30/12
Read Re: Please ignore 4 tables in last post - here are the tables WITHOUT
three clerical errors.
Ray Koopman
5/30/12
Read Re: Please ignore 4 tables in last post - here are the tables WITHOUT
three clerical errors.
Halitsky
5/30/12
Read ln(c/u) on ln(c/L) may not be "weak" but rather 2-3 times stronger
than it should be for study group data.
Halitsky
5/26/12
Read Here are the t-test results for e, c, and u themselves (for a1 fold,
S63 and S60 groups)
Halitsky
5/26/12
Read Two questions re actual c,e,u, expected ln(c/u), ln(c/e), and actual
lnc/u, ln(c/e)
Halitsky
5/26/12
Read Re: Two questions re actual c,e,u, expected ln(c/u), ln(c/e), and
actual lnc/u, ln(c/e)
Ray Koopman
5/26/12
Read Re: Two questions re actual c,e,u, expected ln(c/u), ln(c/e), and
actual lnc/u, ln(c/e)
Halitsky
5/27/12
Read Re: Two questions re actual c,e,u, expected ln(c/u), ln(c/e), and
actual lnc/u, ln(c/e)
Ray Koopman
5/25/12
Read Re: New queston: need your advice on some "study group vs study
group" t-test results
Halitsky
5/25/12
Read Re: New queston: need your advice on some "study group vs study
group" t-test results
Halitsky
5/25/12
Read T-tests of ln(c/e), ln(c/u), ln(c/L), e, u, and c for STudy Groups
(63) and (60)
Halitsky
5/20/12
Read Re: Should I obtain "R" to do Bartlett's (or Fligner's)
homoscedasticity tests on the driver correlations?
Ray Koopman
5/20/12
Read Re: Should I obtain "R" to do Bartlett's (or Fligner's)
homoscedasticity tests on the driver correlations?
Halitsky
5/20/12
Read Re: Should I obtain "R" to do Bartlett's (or Fligner's)
homoscedasticity tests on the driver correlations?
Ray Koopman
5/20/12
Read Re: Should I obtain "R" to do Bartlett's (or Fligner's)
homoscedasticity tests on the driver correlations?
Halitsky
5/15/12
Read Re: Means and SD's for "u" (representation level) in control & study
groups for all six folds
Ray Koopman
5/15/12
Read Re: Means and SD's for "u" (representation level) in control & study
groups for all six folds
Halitsky

Point your RSS reader here for a feed of the latest messages in this topic.

[Privacy Policy] [Terms of Use]

© Drexel University 1994-2014. All Rights Reserved.
The Math Forum is a research and educational enterprise of the Drexel University School of Education.