103,138 reputation
9186364
bio website quantdec.com
location Northeastern US
age 15
visits member for 4 years, 7 months
seen 2 hours ago

Consultant (environmental and spatial stats a specialty), expert witness, and teacher. I can be reached through (outdated but still valid) links posted on my web site.

Twitter: @WilliamAHuber // ASA-P website: http://amstatphilly.org/


Why waste time learning, when ignorance is instantaneous?

--T(iger) Hobbes.

For any complex problem there is a simple solution. And it's always wrong.

--[Mis?]attributed to H.L. Mencken by Dava Sobel, Longitude.


14h
reviewed Close i have problem of my sampling based on secondary data
14h
reviewed Reopen Dissertation-results section
14h
comment Conditional probability of continuous variable
When theory and intuition diverge, one cure is to change the theory. But first you should try to understand the discrepancy. All that "theory" amounts to, in the end, is drawing ineluctable conclusions from a set of axioms that are taken to be "intuitively obvious" or incontrovertibly demonstrated. When one of those conclusions seems intuitively false, then we should seriously consider that our intuition is inconsistent. Indeed, that's the normal state of affairs for human beings. Thus our default reaction should first be to recalibrate our intuition and not be hasty to abandon the theory.
1d
reviewed Close Statistical test to check if mean of two groups are “significantly” same
1d
reviewed No Action Needed daily data series that exist seasonal component
1d
comment Different types of dependent observations
The question is much better now--thank you for the edits--but it still seems problematic, due to the assertion that 50 observations collected from one person would be "dependent." How would that be? Eg,if they were heights, then in most cases an excellent model would be that they can be viewed as independent random variables, with all variation reflecting tiny diurnal changes (which do occur!) together with measurement error. It is always difficult to know how to respond to any question that asserts things that are implausible, vague, contrafactual, or subject to conflicting interpretations.
1d
reviewed Leave Closed Test if two means are significantly different in R
1d
comment Conditional probability of continuous variable
That is a much more interesting argument. I want to give it some thought, because densities are tricky to work with. For instance, if $5$ and $6$ had been replaced by $0$ and $10$, the answer might be different. And if the uniform density on $[0,10]$ had been replaced by an equal mixture of uniform densities on $[0,5)$ and $(5,10]$, some delicate reasoning would be needed--even though the underlying distribution in both cases is exactly the same.
1d
reviewed Close optimization - cost function matlab
1d
comment Conditional probability of continuous variable
Sure it does: there are plenty of ways to transform one continuous distribution to another that swap two values. Actually, your "flipping" didn't even preserve the original distribution. (It changed its support altogether.) So it would appear that all you're doing is replacing one distribution by another one. There doesn't seem to be any principle operating here at all.
1d
comment Conditional probability of continuous variable
Thank you for a thoughtful post. I, for one, seriously doubt the "principle of indifference" will ever be mainstream, because it is not workable. Your argument falls apart when the underlying values are re-expressed. The uniform distribution on $[0,10]$ might thereby become, say, a Cauchy distribution, $5$ could become $0$, and $6$ become $\sqrt{1-\frac{2}{\sqrt{5}}}$. Your "principle of indifference" now produces a completely different answer. (I used the probability transforms to work out this example.)
1d
comment Test if two means are significantly different in R
Welcome to our site. It is highly likely that an elementary question like this already has good answers. Please consider conducting searches on paired t test and p-value.
1d
comment Different types of dependent observations
In order to understand this question I think we will need some explanation of what you mean by "type" of dependent observations, especially since your characterization of "dependent" is extremely broad. Could you edit this post to elaborate on that?
1d
revised Determining whether a successive instances are conditional or independent
edited tags
1d
comment Determining whether a successive instances are conditional or independent
+1 This pithy account covers a lot of ground nicely.
1d
comment Sampling massive dataset
These questions, as stated, are easy to answer: the smallest possible sample consists of one row. If it is selected randomly it is also representative of the population in every dimension. You really need to provide more criteria--of a quantitative nature--to determine what you want to accomplish with this sample. Until then, we just don't have enough information to provide useful answers.
1d
comment Finding quality of fit for two discrete variables with low statistics
Could you explain what you mean by "low statistics"? Suggesting the use of a $\chi^2$ distribution indicates your data are counts--is that really the case? Exactly how were you able to make a prediction via simulation? How did that work?
1d
comment Degeneracy paradox
@ssdecontrol Understanding one of "$\hat y$" and "$y$" to be the bet and the other to be the outcome, and once you adjust that to account for the cost to play and the fact the payoff is not $1$, you would be in agreement with what I wrote :-). Richard, thank you very much for the clarifying edits (+1).
1d
comment Degeneracy paradox
The loss function is the difference between the return on the bet and the cost to play. But your example helps me re-interpret your first line: by "distribution" you seem to refer to the distribution of the outcome (rather than of the loss itself). I was confused because--since the expectation often will not be a whole number--it's difficult to imagine how anyone could possibly "bet on" an expected value!
1d
comment Fitting a variogram model with the pairwise distance matrix supplied
This is quite clever--and without changing the code, is likely the best possible approach if it can be made to work. But almost all variography software accepts points of limited dimensions, typically less than $n=2$ or $n=3$; never more (as far as I am aware). Thus you effectively have to reconstruct the original point configuration (up to a Euclidean motion) from the distance matrix.