<font size="2"><span style="font-family: courier new,courier,monospace;">Hi all,</span><br style="font-family: courier new,courier,monospace;">
<br style="font-family: courier new,courier,monospace;">
<span style="font-family: courier new,courier,monospace;">"are the
signature files saved to the current directory, or in a folder off of
the current directory; then if you change directory before trying the
next module it won't be able to find them?"</span><br style="font-family: courier new,courier,monospace;">
<br style="font-family: courier new,courier,monospace;">
<span style="font-family: courier new,courier,monospace;">Whichever
directory I'm working in, the signature files are always saved under
the image subgroup. If the raster maps I am analyzing are in group
"testi" and subgroup "alatesti", the signature files will be saved in a
directory called "sig" under my subgroup "alatesti". That is, they will
be in /paula/groups/testi/subgroups/alatesti/sig/ </span><br style="font-family: courier new,courier,monospace;">
<br style="font-family: courier new,courier,monospace;">
<span style="font-family: courier new,courier,monospace;">This is
always true, no matter if I create the signatures with i.cluster, or
r.digit & i.gensigset, or i.class. I've tried all these. And the
rights to the signature file are always the same. i.maxlik can access
the </span><br style="font-family: courier new,courier,monospace;">
<span style="font-family: courier new,courier,monospace;">signature files created with i.cluster and i.gensigset, just not the ones created with i.class.</span></font><font size="2"><span style="font-family: courier new,courier,monospace;"> And even if I work in that directory (groups/testi/subgroups/alatesti/sig), I still can't access the signature. </span></font><br style="font-family: courier new,courier,monospace;">
<font size="2"><span style="font-family: courier new,courier,monospace;"></span><br style="font-family: courier new,courier,monospace;">
<span style="font-family: courier new,courier,monospace;">"can you
recreate the error with the spearfish for new NC sample datasets? if so
can you post all commands and messages from that process?"</span><br style="font-family: courier new,courier,monospace;">
<br style="font-family: courier new,courier,monospace;">
<span style="font-family: courier new,courier,monospace;">I strongly suspect I will have the same problem even with the Spearfish
data set, because it occurs no matter which data I am using. I will try
it next, however. Thanks for the tip! The commands I'm using are just: </span><br style="font-family: courier new,courier,monospace;">
<br style="font-family: courier new,courier,monospace;">
<span style="font-family: courier new,courier,monospace;">> i.class map=varikuva group=testi subgroup=alatesti outsig=signature</span><br style="font-family: courier new,courier,monospace;">
<span style="font-family: courier new,courier,monospace;"><br>
-Then there's the whole choosing and analyzing the test regions in the graphic window. When that's finished, I say:</span><br style="font-family: courier new,courier,monospace;">
<span style="font-family: courier new,courier,monospace;"><br>
> i.maxlik group=testi subgroup=alatesti sigfile=signature class=resultmap</span><br style="font-family: courier new,courier,monospace;">
<span style="font-family: courier new,courier,monospace;"><br>
And I get the error message:</span><br style="font-family: courier new,courier,monospace;">
<span style="font-family: courier new,courier,monospace;"><br>
ERROR: Can't read signature file [signature].<br>
<br>
Still, the signature file IS created, and I can see it sitting and lurking there. I just can't use it.<br style="font-family: courier new,courier,monospace;">
</span><br style="font-family: courier new,courier,monospace;">
</font><font style="font-family: courier new,courier,monospace;" face="Courier New,Courier,monospace" size="3"><font size="2">"try removing the @mapset part from the group names in the GUI window."<br>
<br>
I am calling i.class from the command line, and then I'm calling
i.maxlik from the command line as well. So, normally I don't open the
GUI for this. If I do open the GUI window for the i.maxlik, and try
from there, and make sure that the @mapset doesn't appear, I still get
the same error message.<br>
<br>
Thanks for the messages and the suggestions!<br>
<br>
Paula<br>
</font></font>