FAQ

Navigation

This wiki

This page

- How the isotropic volume size of the DARTEL can be changed?
There is no option for that in the matlabbatch interface. However it can be changed in the code of the SPM8 toolbox. A change change on line 211 of ../spm8/toolbox/Seg/spm_preproc_write8.m from:
vx = 1.5;
to
vx = 1.0;
should change the default voxel size to 1mm.

- How to find appropriate data?

lifmetstor1: /shared/projects/adni_test/spm_test/List and Filter.xlsm

This is a Excel file with a macro that filters the data according to a number of criteria.

- How to transfer the data from lifmetstor1 to your machine?

- usage of ftp in a MATLAB script

- usage of wget on Windows

- How to check the homogeneity of the selected data?

Use VBM8 tools (see wiki)

- How to easily create a batch to create a template?

Use the sample file in /shared/projects/adni_test/spm_test/

- How to get a dot product matrix and what is it good for?

The dot product matrix is the result of the pipeline. It is always prefixed by a "dp_"

It contains the computed kernels for a linear SVM.

- How to use the SVM?

See this wiki.

- Where to find more information?

In this wiki.

- Does it make a big difference if I scale each attribute to [0,1] instead of [-1,1]?

For the linear scaling method, if the RBF kernel is used and parameter selection is conducted, there is no difference. Assume Mi and mi are respectively the maximal and minimal values of the ith attribute. Scaling to [0,1] means

x'=(x-mi)/(Mi-mi)

For [-1,1],

x''=2(x-mi)/(Mi-mi)-1.

In the RBF kernel,

x'-y'=(x-y)/(Mi-mi), x''-y''=2(x-y)/(Mi-mi).

Hence, using (C,g) on the [0,1]-scaled data is the same as (C,g/2) on the [-1,1]-scaled data.

Though the performance is the same, the computational time may be different. For data with many zero entries, [0,1]-scaling keeps the sparsity of input data and hence may save the time.

Question and answer taken from: http://www.csie.ntu.edu.tw/~cjlin/libsvm/faq.html#/Q3:_Data_preparation