| |
 |
|
|
Science Forum Index » Statistics - Math Forum » Warning
Page 1 of 1
|
| Author |
Message |
| Luis A. Afonso |
 Posted: Fri May 02, 2008 1:11 am |
|
|
|
Guest
|
Warning
Specially directed to out co-Readers in order that DO NOT BELIEVE WHAT the two-imbeciles in one are repeatedly said concerning p-values are not possible to get by Monte - Carlo
The IMBECILE, is Jack Tomsky who, by times uses the pen-name John Smith.
From the WEB
www.technion.ac.il/docs/sas/stat/chap28/sect28.htm
Exact Statistics
Exact statistics can be useful in situations where the asymptotic assumptions are not met, and so the asymptotic p-values are not close approximations for the true p-values. Standard asymptotic methods involve the assumption that the test statistic follows a particular distribution when the sample size is sufficiently large. When the sample size is not large, asymptotic results may not be valid, with the asymptotic p-values differing perhaps substantially from the exact p-values. Asymptotic results may also be unreliable when the distribution of the data is sparse, skewed, or heavily tied. Refer to Agresti (1996) and Bishop, Fienberg, and Holland (1975). Exact computations are based on the statistical theory of exact conditional inference for contingency tables, reviewed by Agresti (1992).
Under, WEB,
_____p- values by Monte Carlo
. . every one can found a lot of references in this direction (to can read, is enough).
LuÃs Amaral Afonso (The Moderator Destroyer) |
|
|
| Back to top |
|
| John Smith |
| Posted: Fri May 02, 2008 7:43 am |
|
|
|
Guest
|
Adumo pronounces on p-values.
Adumbo, do you mean real p-values or the p-values that you mistakenly think are significance levels????
This is hilarious, coming from someone who can't tell a p-value from a signficance level!!!!
AFONSO WROTE:
As I found
__________1 – alpha =______0.99868 (equal
variances)
__________1 – alpha =______0.99832 (different)
I immediately got the SIGNIFICANT LEVELS:
0.00132, 0.00168 (0.13%, 0.17%) (***Brilliant
isn’t?***)
__________
Luis
TOMSKY WROTE:
These are p-values, not significance levels.
Jack (moderator of the Math Forum) |
|
|
| Back to top |
|
| Luis A. Afonso |
| Posted: Fri May 02, 2008 10:31 am |
|
|
|
Guest
|
Jack Tomsky (under the pseudonymous of John Smith, or vice –versa)
I explained yet (a thousand of times) how the true statisticians (not you, of course) achieve DECISIONS using Monte Carlo:
Supposing a ONE RIGHT TAIL HYPOTHESES TEST
___a) real data (one or two samples) is drawn from Nature. Chosen Alpha = 5% for example as the Type I error probability. Using the adequate Test Statistics W one calculate the respective value : be W0,
___b) By simulation (supposing H0 true) one obtain the EMPIRICAL DISTRIBUTION FUNCTION of the test statistics. Let be Ws (s by simulated) the point limiting the 0.95 fractile.
___c) DECISION
H0 is rejected only if Ws>W0, what is the same to have prob(W>=Ws)<= Alpha
Fail to reject H0, otherwise. In fact Ws<=W0 and prob(W>=Ws)> alpha.
*********************
The situation of FAIL TO REJECT H0
_____W0____Ws,alpha->__________
Reject H0
________Ws,alpha->________W0_________
OTHER PROCEDURE
__a) From the simulated data one find out the proportion of W such that are greater than W0, if is greater than the pre-chosen Alpha fail to reject H0, reject otherwise.
This method (contrarily to the former) demands that one must perform simulations every time NRW data is tested. Not ADVISABLE.
LuÃs Amaral Afonso (The Moderator Destroyer) |
|
|
| Back to top |
|
| Jack Tomsky |
| Posted: Fri May 02, 2008 1:34 pm |
|
|
|
Guest
|
Quote: Jack Tomsky (under the pseudonymous of John Smith, or
vice –versa)
I explained yet (a thousand of times) how the true
statisticians (not you, of course) achieve DECISIONS
using Monte Carlo:
Supposing a ONE RIGHT TAIL HYPOTHESES TEST
___a) real data (one or two samples) is drawn from
Nature. Chosen Alpha = 5% for example as the Type I
error probability. Using the adequate Test
Statistics W one calculate the respective value : be
W0,
___b) By simulation (supposing H0 true) one obtain
the EMPIRICAL DISTRIBUTION FUNCTION of the test
statistics. Let be Ws (s by simulated) the point
limiting the 0.95 fractile.
___c) DECISION
H0 is rejected only if Ws>W0, what is the same to
have prob(W>=Ws)<= Alpha
Fail to reject H0, otherwise. In fact Ws<=W0 and
prob(W>=Ws)> alpha.
LuÃs Amaral Afonso (The Moderator Destroyer)
Under the Afonso 4.0 version of hypothesis testing, we reject Ho if the sample W is less than Ws, which has a probability of 1-alpha = 0.95 under the null hypothesis.
Similarly, we fail to reject Ho if W is greater or equal than Ws, which has a probability of alpha = 0.05 under the null hypothesis.
This, being the reverse of what statisticians would do, I recommend that no one ever use Afonso 4.0.
Jack (moderator) |
|
|
| Back to top |
|
| licas_@hotmail.com |
| Posted: Fri May 02, 2008 2:47 pm |
|
|
|
Guest
|
Jack Tomsky escreveu:
Quote: Jack Tomsky (under the pseudonymous of John Smith, or
vice –versa)
I explained yet (a thousand of times) how the true
statisticians (not you, of course) achieve DECISIONS
using Monte Carlo:
Supposing a ONE RIGHT TAIL HYPOTHESES TEST
___a) real data (one or two samples) is drawn from
Nature. Chosen Alpha = 5% for example as the Type I
error probability. Using the adequate Test
Statistics W one calculate the respective value : be
W0,
___b) By simulation (supposing H0 true) one obtain
the EMPIRICAL DISTRIBUTION FUNCTION of the test
statistics. Let be Ws (s by simulated) the point
limiting the 0.95 fractile.
___c) DECISION
H0 is rejected only if Ws>W0, what is the same to
have prob(W>=Ws)<= Alpha
Fail to reject H0, otherwise. In fact Ws<=W0 and
prob(W>=Ws)> alpha.
Luís Amaral Afonso (The Moderator Destroyer)
Under the Afonso 4.0 version of hypothesis testing, we reject Ho if the sample W is less than Ws, which has a probability of 1-alpha = 0.95 under the null hypothesis.
Similarly, we fail to reject Ho if W is greater or equal than Ws, which has a probability of alpha = 0.05 under the null hypothesis.
This, being the reverse of what statisticians would do, I recommend that no one ever use Afonso 4.0.
Jack (moderator)
THIS GUY IS DRUNK OR DRUGGIE
Luis Amaral Afonso (The Moderator Destroyer) |
|
|
| Back to top |
|
| Jack Tomsky |
| Posted: Fri May 02, 2008 3:32 pm |
|
|
|
Guest
|
Quote:
Jack Tomsky escreveu:
Jack Tomsky (under the pseudonymous of John
Smith, or
vice –versa)
I explained yet (a thousand of times) how the
true
statisticians (not you, of course) achieve
DECISIONS
using Monte Carlo:
Supposing a ONE RIGHT TAIL HYPOTHESES TEST
___a) real data (one or two samples) is drawn
from
Nature. Chosen Alpha = 5% for example as the Type
I
error probability. Using the adequate Test
Statistics W one calculate the respective value :
be
W0,
___b) By simulation (supposing H0 true) one
obtain
the EMPIRICAL DISTRIBUTION FUNCTION of the test
statistics. Let be Ws (s by simulated) the point
limiting the 0.95 fractile.
___c) DECISION
H0 is rejected only if Ws>W0, what is the same to
have prob(W>=Ws)<= Alpha
Fail to reject H0, otherwise. In fact Ws<=W0
and
prob(W>=Ws)> alpha.
LuÃs Amaral Afonso (The Moderator Destroyer)
Under the Afonso 4.0 version of hypothesis testing,
we reject Ho if the sample W is less than Ws, which
has a probability of 1-alpha = 0.95 under the null
hypothesis.
Similarly, we fail to reject Ho if W is greater or
equal than Ws, which has a probability of alpha =
0.05 under the null hypothesis.
This, being the reverse of what statisticians would
do, I recommend that no one ever use Afonso 4.0.
Jack (moderator)
THIS GUY IS DRUNK OR DRUGGIE
Luis Amaral Afonso (The Moderator Destroyer)
Afonso is mad at me because I pointed out that his test which has a chosen significance level of 5% has a probability of rejecting the null hypothesis of 95%. Since 95% is greater than 5%, the Afonso test is inadmissible and should never be used.
Jack (moderator) |
|
|
| Back to top |
|
| Luis A. Afonso |
| Posted: Fri May 02, 2008 3:37 pm |
|
|
|
Guest
|
I wrote:
*********************
The situation of FAIL TO REJECT H0
_____W0____Ws,alpha->__________
Reject H0
________Ws,alpha->________W0_________
Even an APE understand that Ws is the critical value (for the chosen alpha) and this diagram is clear: If the real world sample shows a value W0 that is LESS THAN (or equal to) Ws, fail to reject ; if , on contrary W0 > Ws = critical value THEN Reject the Null Hypotheses.
Wickedly Jack did not how this part of my post. One do not expect otherwise from this disgusting guy.
It’s each time more evident that his understanding and knowledge in what concerns Statistical Tests by simulation is HORRIFYING.
Luis Amaral Afonso (The Moderator Destroyer)
*****************************
Home work to Jack Tomsky (sometimes John Smith)
The worse consequence is that they not get that must re-think what Statistics is. Jack Tomsky is a one more eloquent example.. If one has not full conscience that Probability is the realm of RANDOMNESS, Jack Tomsky would never was caught in such BRUNDLES as, for example:
__a)____The Null Hypotheses can be proved true
__b)____H0: m=0 if rejected means that the parameter m is truly ZERO.
__c)____The of asymptotical values use are always correct whatever the sizes evolved in the Hypotheses Tests,
__d)___The dependence critical values on sample sizes (for a given alpha) should be monotonous increasing,
__e)___The sample statistics (even if parameters are evaluated fro data) must follow a given Distribution (cf. Jarque - Bera Test and the Chi-squared, two degrees of freedom),
__f)____The Empirical Distribution Function of a Test Statistics (because is obtained through a sample) is not able to provide critical Values.
*************************************
Luis Amaral Afonso (The |
|
|
| Back to top |
|
| Jack Tomsky |
| Posted: Fri May 02, 2008 4:21 pm |
|
|
|
Guest
|
Quote: I wrote:
*********************
The situation of FAIL TO REJECT H0
_____W0____Ws,alpha->__________
Reject H0
________Ws,alpha->________W0_________
Even an APE understand that Ws is the critical value
(for the chosen alpha) and this diagram is clear: If
the real world sample shows a value W0 that is LESS
THAN (or equal to) Ws, fail to reject ; if , on
n contrary W0 > Ws = critical value THEN Reject the
Null Hypotheses.
Wickedly Jack did not how this part of my post. One
do not expect otherwise from this disgusting guy.
It’s each time more evident that his understanding
and knowledge in what concerns Statistical Tests by
simulation is HORRIFYING.
Luis Amaral Afonso (The Moderator Destroyer)
*****************************
Home work to Jack Tomsky (sometimes John Smith)
The worse consequence is that they not get that must
re-think what Statistics is. Jack Tomsky is a one
more eloquent example.. If one has not full
conscience that Probability is the realm of
RANDOMNESS, Jack Tomsky would never was caught in
such BRUNDLES as, for example:
__a)____The Null Hypotheses can be proved true
__b)____H0: m=0 if rejected means that the parameter
m is truly ZERO.
__c)____The of asymptotical values use are always
correct whatever the sizes evolved in the Hypotheses
Tests,
__d)___The dependence critical values on sample sizes
(for a given alpha) should be monotonous increasing,
__e)___The sample statistics (even if parameters are
evaluated fro data) must follow a given Distribution
(cf. Jarque - Bera Test and the Chi-squared, two
degrees of freedom),
__f)____The Empirical Distribution Function of a Test
Statistics (because is obtained through a sample) is
not able to provide critical Values.
*************************************
Luis Amaral Afonso (The
This is what Afonso wrote in recommending his test where the probability of rejecting the null hypothesis when it's true exceeds the significance level. He does not believe in being mathematically correct since he considers mathematicians to be a "pitiable gang".
"___c) DECISION
H0 is rejected only if Ws>W0, what is the same to have prob(W>=Ws)<= Alpha
Fail to reject H0, otherwise. In fact Ws<=W0 and prob(W>=Ws)> alpha."
Jack (moderator) |
|
|
| Back to top |
|
| |
|
Page 1 of 1
All times are GMT - 5 Hours
The time now is Sun Jul 06, 2008 8:02 pm
|
|
