11,692 reputation
31945
bio website
location England
age
visits member for 1 year, 11 months
seen 2 days ago

2d
revised How should I model interactions between explanatory variables when one of them may have quadratic and cubic terms?
edited tags
2d
revised How to calculate $P (|X − Y | ≤ 1/6)$?
edited tags
2d
comment How to get probability response from LDA classifier in R?
Read the manual: ?predict.lda. The posterior probabilities certainly shouldn't be negative - are you looking at the right vectors in res?
2d
reviewed Close How to get probability response from LDA classifier in R?
2d
reviewed Close How to plot model with forecasts in R?
2d
reviewed Close Please help me about Dunnett's test in R
Nov
20
comment Logistic Regression - Error Term and its Distribution
@whuber: I've corrected my answer wrt (3), which wasn't well thought through; but still puzzled about in what sense (2) might be right.
Nov
20
revised Logistic Regression - Error Term and its Distribution
added 123 characters in body
Nov
20
answered Multiple Regression with Categorical Predictor Variables of More than Two Levels
Nov
20
revised What does the Argument “type” in VAR() - function do?
deleted 8 characters in body
Nov
20
comment What does the Argument “type” in VAR() - function do?
Thanks; that's better
Nov
20
comment My naive bayes classifier doesn't show probabilities
Welcome to Cross Validated! It's not really the place, though, to ask questions that are solely about how to use software: try the Weka help resources here.
Nov
20
comment Logistic Regression - Error Term and its Distribution
@Glen_b: Might one argue for (2)? I've known people to say it but never to defend it when it's questioned.
Nov
20
revised Logistic Regression - Error Term and its Distribution
added 321 characters in body
Nov
20
comment Error distribution for linear and logistic regression
B_Miner: $\operatorname{Var}(Y|X)=\operatorname{Var}(Y|X=x)$ means the variance of $Y$ conditional on the random variable $X$'s taking an observed value $x$. So it's immaterial whether your predictors are fixed by an experiment or observed in a sample: what @Stat's saying is they're no longer being considered as random variables for the purposes of regression.
Nov
20
revised Error distribution for linear and logistic regression
fixed typo
Nov
20
revised independent replicate experiments yielding contrasting results
edited tags; edited tags
Nov
20
comment Error distribution for linear and logistic regression
NB $e_i=1−p_i$ with probability $p_i$ or $e_i=−p_i$ with probability $1−p_i$ is not a binomial distribution for $e_i$.
Nov
20
revised Multiple Linear Regression Variable Selection
edited tags; edited tags
Nov
20
answered Logistic Regression - Error Term and its Distribution