[R-sig-Geo] A request

[shamsman87 at yahoo.com]

Dear all 
Hi 
I hope you are well, I am Mansour ,Ph.D student of epidemiology in
Tehran university of medical sciences , with the help of some biostatistician,We are going to a run spatio-temporal model on
air pollution data . We are using the SpatioTemporal package.  
(http://cran.r-project.org/web/packages/SpatioTemporal/vignettes/ST_tutorial.pdf). 
we run it, but in estimation part we come across below error (highlighted
) that we cannot fixed it.  
>
est.EBOD.model <- estimate(EBOD.model, x.init, type="f",
hessian.all=TRUE)
Optimisation
using starting value 1/2
N = 13, M = 5
machine precision = 2.22045e-16
The initial X
is infeasible.  Restart with its
projection.
At X0, 2
variables are exactly at the bounds
At
iterate     0  f=        94332  |proj g|=           15
At
iterate    10  f =        12862  |proj g|=        6.1852
At
iterate    20  f =        12861  |proj g|=        1.8072
At
iterate    30  f =        12848  |proj g|=        2.0701
At
iterate    40  f =        12848  |proj g|=       0.40681
 
iterations 41
function
evaluations 54
segments
explored during Cauchy searches 47
BFGS updates
skipped 0
active bounds
at final generalized Cauchy point 2
norm of the
final projected gradient 0.0924766
final function
value 12847.5
 
F = 12847.5
final  value 12847.531525 
converged
 
Error in diag(solve(res[[i]]$hessian)) : 
  error in
evaluating the argument 'x' in selecting a method for function 'diag': Error in
solve.default(res[[i]]$hessian) : 
  Lapack
routine dgesv: system is exactly singular: U[4,4] = 0
I attached our script,
output and sample of data which we used.  The error above seems simple but we can't find
how it occurs. Any help you can give would be greatly appreciated. 
 
Best 
Mansour 
 
 
 
Best Regards, 
Mansour Shamsipour 
Ph.D Candidate in Epidemiology 
Tehran University of Medical Science,(TUMS)  
Mobile: (+98-914) 939 2486 
021-55460184
	[[alternative HTML version deleted]]