Hi,
I am currently trying to create a kernel.pca with a meta-kernel object (from combine.kernels()) using the latest version of mixKernel.
I am having some issue with this function when running the kernel.pca, I am getting the following error :
"Error in names(per.var) <- paste("PC", 1:length(per.var), sep = "") : attempt to set an attribute on NULL"
which when I then try to plot the PCA using the plotIndiv() gives me:
"Error in round(inf, 2) : non-numeric argument to mathematical function"
But when I try to run a pca with a normal pca() function using the meta.kernel$kernel it works (only for one dimension at a time).
What I also do not understand is that even when I try using the TARAoceans data provided from the online tutorial (http://mixomics.org/mixkernel/) and the exact same code I am getting the same error and cannot make a plot either.
Hi, I am currently trying to create a kernel.pca with a meta-kernel object (from combine.kernels()) using the latest version of mixKernel.
I am having some issue with this function when running the kernel.pca, I am getting the following error : "Error in names(per.var) <- paste("PC", 1:length(per.var), sep = "") : attempt to set an attribute on NULL" which when I then try to plot the PCA using the plotIndiv() gives me: "Error in round(inf, 2) : non-numeric argument to mathematical function"
But when I try to run a pca with a normal pca() function using the meta.kernel$kernel it works (only for one dimension at a time).
What I also do not understand is that even when I try using the TARAoceans data provided from the online tutorial (http://mixomics.org/mixkernel/) and the exact same code I am getting the same error and cannot make a plot either.
Have anyone ever came accross this issue before?