Dear FSL experts,
Can you please help me interpret these results from FSL Nets ?
1. arone =
0.1117
RtoZcorrection =
17.4595
arone =
0.1117
RtoZcorrection =
17.6594
2. Here is another result, which is from the attached figure
sorted edge#1 nodes 7,4 value=0.997
What does this mean ? Does this value mean the p-value ?? But sometimes I get values around 0.62 for some nodes; while all the components used in the analysis were above 0.99 p value threshold after benferroni correction.
3. I m also not abe to understand the LDA percentages
[lda_percentages]=nets_lda(netmats3,19, 1) : FLD gives me
lda_percentages =
76.1905
[lda_percentages]=nets_lda(netmats3,19, 3); which is two group T weighting is
lda_percentages =
90.4762
I m unable to understand the practical difference between these results; which one am I supposed to use ?
4. On the other hand
[lda_percentages]=nets_lda(netmats3,19, 8) for libsvm gives me an error because the feature vectors in libsvm are mistakenly getting understood as group-names.
svmstruct = svmtrain([ ones(size(xa,2),1) ; zeros(size(xb,2),1) ], [xa xb]' , '-q');
xa is 342x18 matrix. xb is 342x2 matrix. Can you please check and guide what to correct here ?
5. If I m trying to find the network differences between 2 groups; and given the clusters figure; what should I be able to interpret ? I can understand that these clusters represents how the good components are connected to each other based on the netmats method that I have used. But if I m looking to find network differences in two groups; how should I select the best netmats results on which this cluster figure is based upon ??
Any help will be highly appreciated
Thanks a lot and best,
Qasim