Displaying 3 results from an estimated 3 matches for "lm_object".
Did you mean:
vm_object
2010 Sep 23
1
Newey West and Singular Matrix + library(sandwich)
....)
>
> 1. There is no assert() function in base R.
> 2. The error message of se.neweywest() refers to se.white.
> 3. A much more flexible and powerful solution of this is included
> ? in package "sandwich", see the vignette for details. The code
> ? ? sqrt(diag(NeweyWest(lm_object, lag = 0, prewhite = 0)))
> ? replicates se.neweywest(lm_object) but has the following advantages:
> ? it also does automatic bandwidth selection, it does not require setting
> ? "x = TRUE", it incorporates other kernel weighting functions, supports
> ? prewhitening etc.
>...
2010 Sep 22
1
Newey West and Singular Matrix
dear R experts: ?I am writing my own little newey-west standard error
function, with heteroskedasticity and arbitrary x period
autocorrelation corrections. ?including my function in this post here
may help others searching for something similar. it is working quite
well, except on occasion, it complains that
Error in solve.default(crossprod(x.na.omitted, x.na.omitted)) :
system is
2011 Jun 23
0
R-squared values for multiple linear regression with a matrix of multiple response variables
...rix Y that is explained by
the predictor matrix X.
I have been using manova(Y ~ X) to assess the significance of the
linear model. I am also using lm(Y ~ X) or lm(cbind(y1, y2, ...) ~ x1
+ x2 + x3 +....) but these seem to fit separate multiple linear models
to each response variable, i.e., summary(lm_object) would return a
list of regression summaries for each response variable.
I would actually like to fit a model on the two matrices with one as
the response and the other as the predictor, and compute an R2 value
of the correlation between the two matrices. Is there a built-in
function in R that doe...