HAKAN DEMIRTAS wrote:
> I can't seem to get computationally stable estimates for the following system:
>
> Y=a+bX+cX^2+dX^3, where X~N(0,1). (Y is expressed as a linear combination of
> the first three powers of a standard normal variable.) Assuming that E(Y)=0
> and Var(Y)=1, one can obtain the
I can't seem to get computationally stable estimates for the following system:
Y=a+bX+cX^2+dX^3, where X~N(0,1). (Y is expressed as a linear combination of
the first three powers of a standard normal variable.) Assuming that E(Y)=0 and
Var(Y)=1, one can obtain the following equations after tedio
Prof Brian Ripley <[EMAIL PROTECTED]> writes:
> Those functions are messy (StringFromRaw was always private), and are
> really just part of coerceVector and asInteger etc (which are the public
> interface).
>
> What RdbiPgSql does is not currently correct. It has
>
> PROTECT(rowNames = allocV
The S4 type introduced for testing last month is now the standard in the
r-devel (2.4) version of R.
Note that older binary objects, including class definitions, if restored
will not be consistent with the new type. They will probably work most
of the time, but should be regenerated from sourc
>From the help page (?unique)
Value:
For a vector, an object of the same type of 'x', but with only one
copy of each duplicated element. No attributes are copied (so the
result has no names).
Please take your own advice and `learn by doing as you are asked in the
posting guide'
