Displaying 1 result from an estimated 1 matches for "infomatrix_inv".
2008 Mar 05
2
matrix inversion using solve() and matrices containing large/small values
...dosis^macht)
> sigmadosis = c(1:7)
> for(i in 1:7){
sigmadosis[i]=ifelse(sigerr[i]<2.257786084*10^(-4),2.257786084*10^(-4),sigerr[i])
}
> omega = diag(sigmadosis)
> infomatrix = dfdbtrans%*%omega%*%dfdb
**********************
I need the inverse of this information matrix, and
> infomatrix_inv = solve(infomatrix, tol = 10^(-43))
does not deliver adequate results (matrixproduct of infomatrix_inv and
infomatrix is not 1). Regular use of solve() delivers the error 'system is
computationally singular: reciprocal condition number: 2.949.10^(-41)'
So I went over to an eigendecomposi...