Dear
R-users,
I was
wondering if someone could give me some advices on the following problem.
I tried to
apply moran’s test to a small dataset and couldn’t succeed, here is the error
message:
mor <- moran.test(x, res2)
Avis dans
moran.test(x, res2) :
Out-of-range p-value: reconsider test arguments
mor
Moran's I test under randomization
data: x
weights: res2
Moran I statistic standard deviate = NaN, p-value = NA
alternative hypothesis: greater
sample estimates:
Moran I statistic Expectation Variance
-0.04166667 -0.04166667 0.00000000
It seems that I have a problem computing the
standard deviation, variance and/or p-value.
As a brief
overview of what are x and res2: I have a square sample plot on the field
divided
into 5 lines / 5 columns and one value per “sub-plot”, so a dataframe of 25
values (25 cells and 1 value per cell…).
The parameter x is a numeric vector containing those 25 values:
str(x)
num [1:25]
22920 19546 16170 21387 9499 ...
To obtain
my spatial weight matrix, I used 2 columns (X & Y) representing the
coordinates of the center of each sub-plot (20x20m, so 10:10, 10:30, 10:50,
etc.), transformed it into a neighborhood matrix of class “nb” with dnearneigh()
function, then into a object of class “listw” with nb2listw() function (with
style=”B”). I end up with “res2” :
res2
Characteristics of weights list
object:
Neighbour list object:
Number of regions: 25
Number of nonzero links: 600
Percentage nonzero weights: 96
Average number of links: 24
Weights style: B
Weights constants summary:
n nn S0
S1 S2
B 25 625 600 1200 57600
Ok so it seems quiet straightforward
to me and I cannot figure out from where my problem is coming !!! any idea ?
with regards
[[alternative HTML version deleted]]
