92,231 reputation
8158323
bio website quantdec.com
location Northeastern US
age 14
visits member for 4 years, 3 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 I can be reached through (outdated but still valid) links posted on my web site.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.


3h
comment Conceptual question on correlation matlab and implication
The beginning of this post is very confusing: the formula you give is for correlation; there is nothing in this setting that gives any meaning to "autocorrelation"; in neither case will correlations (or autocorrelations) produce "mean values." Thus it's impossible to determine what you are asking and equally impossible to respond to a question about what should be used.
3h
comment Data transformation using copulas
@Horst Or maybe the intention is to make the marginals Normal. Or maybe both (in some sense). It's unclear.
3h
comment Kolmogorov-Smirnov and Hellinger distances
Could you explain what you mean by "no correlation" in a context where only two numbers are available for comparison? (Correlation ordinarily describes a property of some collection of ordered pairs of numbers, which could be graphed with a scatterplot.)
3h
comment Regression - Censored Data
(1) Where is the censoring you describe? You haven't provided any information about this except to state the data do have some censoring. We don't even know whether the censoring was above, below, or within intervals! Are you sure these data are actually censored in the usual statistical meaning of the term? (2) What is the nature of this heteroscedasticity you mention and what is the evidence for it? The weights you use are very strong for data that act like test scores.
4h
comment Data transformation using copulas
What exactly do you mean by "more Normal"? Are you saying the marginals of your data do not appear to be Normal and you would like them to be? Or are you referring to the correlation structure of your data (and if so, how does one compare the "normalness" of different correlation matrices)?
4h
comment What is a “log-F” distribution?
Since the log of the statistic has an F distribution, then why not base your computation of the p-value on the log of the statistic?
4h
comment What is the difference between Harris Corner Detection and SUSAN Corner Detection?
This question might be on topic at Signal Processing
5h
revised Quick Exploratory Analysis of Categorical Data
edited tags
5h
comment Square root of number of counts, or standard deviation of the mean?
What is the possibility that the underlying intensity could vary appreciably among the three replications of the experiment? Does "$n$" refer to the total of the three counts, the mean, or the individual counts? Are the counts measured accurately or is there some measurement error? Are you using raw counts or have you subtracted a background estimate?
5h
comment Need clarification on a passage. (Autism Prevalence)
What exactly are the relationships among the "social/communicative" distinction at the end and the "mental retardation/high-functioning" classifications discussed earlier?
5h
reviewed Looks OK Kernel matrix is a covariance function
6h
comment Best linear unbiased estimator
It would be much simpler and less confusing, then, not to mention time at all. But that leaves us puzzling over how you come up with $\hat\sigma^2_{i,t}$, since it seems you would have only one observation for each $i$.
6h
revised What is a “log-F” distribution?
added 53 characters in body
6h
revised What is a “log-F” distribution?
added 6 characters in body; edited title
6h
comment Best linear unbiased estimator
Thank you. It looks like the answer must depend on how you estimate the $\hat\beta_k$, so you ought to explain that, too. What you are doing remains a murky: although you state you are estimating a "fitted variance ... at each time $t$", your formula for it makes no reference to $t$ at all. Does time actually play any role here or are you just performing a sequence of independent estimates, one for each time? Regardless, how do you estimate $\hat\sigma^2_{i,t}$ for a single stock $i$ at a single time $t$?
6h
revised What is a “log-F” distribution?
added 541 characters in body
6h
comment Best linear unbiased estimator
What exactly is a "cross-section average of variance"? Whatever it is, is it assumed to be known or is it estimated from the same data? What information do you have about correlations among the returns?
8h
reviewed Reviewed Negative Binomial Anscombe Residuals
8h
reviewed Reviewed How to reconstruct the data?
8h
reviewed Reviewed Density of Argmax of a Gaussian Process