11,013 reputation
31945
bio website
location England
age
visits member for 1 year, 11 months
seen 1 hour ago

6h
comment How does MATLAB compute the correlation coefficient for robust nonlinear least squares?
E.g. Koener & Machado (1999), "Goodness of fit & related inference processes for quantile regression", JASA, 94, 448).
7h
revised How does MATLAB compute the correlation coefficient for robust nonlinear least squares?
expanded "LAR", fixed typos
7h
comment How does MATLAB compute the correlation coefficient for robust nonlinear least squares?
Is the coefficient of determination (according to the standard definition) relevant when you're optimizing by another criterion? The quantile regression function rq in R's quantreg doesn't calculate it at all; perhaps Matlab is giving some more appropriate equivalent.
1d
reviewed No Action Needed What's the difference between correlation and simple linear regression?
1d
comment What's the difference between correlation and simple linear regression?
Welcome to CV! Given that there are so many answers to this question already, do you want to have a look at them & see if yours adds anything new? If you've more to say, you can edit it to do so.
1d
comment What exactly are censored data?
... Your cell count data are indeed censored according to many people's understanding of the term, which is not restricted to time-to-event measurements, because you know everything about each subject except how far below 300 his cell count is; "truncation" here (alternatively "Winsorization") describes the method of analysis, viz the treatment of values below 300 as if they were equal to 300.
1d
comment What exactly are censored data?
There's another use of "truncation": to describe a data-generating process where observations above/below cut-offs are unobtainable. A classic example involves counting the no.eggs found in the nests of a particular bird species, where the species can only be identified from the egg; empty nests could be from any species so the no. zeroes is unknown. If the no. eggs follows a Poisson distribution, the egg counts from non-empty nests follow a truncated Poisson. So truncation produces missing data according to a specific well defined mechanism.
1d
comment For Y~B(11, 0.3) what is P(|Y-5| >= 3)?
Welcome to Cross Validated! Please note this is a self-study question. The other answers are deliberately incomplete.
1d
revised How should I evaluate the expectation of the ratio of two random variables?
fixed typos
1d
revised How to solve hat check problem using R?
edited tags
1d
comment How to solve hat check problem using R?
Please don't expect us to do your homework for you. Try to do it yourself & if you get stuck post a question explaining where either here (if it's to do with statistics) or on Stack Overflow (if it's more to do with R programming); then someone will give you a hint.
1d
reviewed Leave Closed Systematic Review (maybe)
1d
reviewed Approve suggested edit on How to solve hat check problem using R?
1d
reviewed Approve suggested edit on time-series analysis Vs statistical signal processing
1d
comment Using MySQL or r, or both?
(1) If the data-set is too large for memory (or if you want it in a database for other reasons) then you can smooth the workflow by sending SQL commands from R using e.g. RMySQL to retrieve subsets as data frames. (2) Otherwise you can read the data into R & perform the data manipulation within R using either (a) SQL syntax see sqldf, or (b) native R commands for subsetting, joining &c., or (c) other R packages like plyr or reshape2
1d
comment Simplifying variable effects in a GLM in R
And for the record, you could do all those things in base R - see the levels, poly, & cut functions, & note that a model object contains an editable $coefficients list. (Though "tweaking" effect estimates is a terrible idea - unless the predictors are orthogonal, changing one alters the best estimate of the others. If you want to incorporate prior beliefs about effect sizes into models you should be asking about packages for Bayesian inference in R.bayessm might be a good one to start with.)
2d
comment Simplifying variable effects in a GLM in R
"Smoothing" ordinal predictors might be done by taking only the first few orthogonal polynomials, or you could go to town on it & use a penalized likelihood approach where the penalty's bigger when differences in coefficients for adjacent categories are larger - see here. My tirade against banding continuous variables might also interest you.
2d
comment Why is permutation test wrong with variance but correct with standard deviation?
A glance at the code shows your test statistics are the difference in sample standard deviations & the difference in sample variances; these don't order the permutations the same way (ratios would).
2d
reviewed Reviewed values in the grouping variable equal “..” !
2d
reviewed Leave Closed Why is permutation test wrong with variance but correct with standard deviation?