Take the 2-minute tour ×
Cross Validated is a question and answer site for people interested in statistics, machine learning, data analysis, data mining, and data visualization. It's 100% free, no registration required.

I'm new to SVM and would like to use it to solve a problem formulated, for example, as follows:

The patterns are four-dimensional vectors $(x_1,x_2,x_3,x_4)$, and the kernel is $K(\bf{x_i}, \bf{x_j})=k1([{x_i}_1, {x_i}_2], [{x_j}_1, {x_j}_2]) \times k2([{x_i}_3, {x_i}_4], [{x_j}_3, {x_j}_4)$

In other words, the kernel is the multiplication of two other kernels, each of which takes a pair of sub-patterns as input.

I came across libSVM and would like to know if the library is able to solve this kind of problem. Thanks.

share|improve this question
    
Do you know what k1 and k2 are, and can you calculate them? If so, then no problem! Any svm library would work! –  Stumpy Joe Pete Oct 19 '12 at 18:15
    
I plan to make k1 a Gaussian kernel and k2 a RBF kernel. According to its tutorial, it seems libSVM only allows you choose one kernel type out of linear, polynomial, rbf and sigmoid. How do I tell the library I want a multiplicative kernel? Please excuse me if this question is too basic, I just quickly went through a book on SVM yesterday –  user11869 Oct 19 '12 at 18:22
    
The webpage mentioned that you can provide a precomputed kernel matrix. I have to go track down the documentation on how to do that, but the basic idea is to compute the kernel matrix yourself and to tell libsvm to use that –  Stumpy Joe Pete Oct 19 '12 at 21:16

1 Answer 1

up vote 1 down vote accepted

From the README file, on the topic of precomputed kernels:

Precomputed Kernels 
===================

Users may precompute kernel values and input them as training and
testing files.  Then libsvm does not need the original
training/testing sets.

Assume there are L training instances x1, ..., xL and. 
Let K(x, y) be the kernel
value of two instances x and y. The input formats
are:

New training instance for xi:

<label> 0:i 1:K(xi,x1) ... L:K(xi,xL) 

New testing instance for any x:

<label> 0:? 1:K(x,x1) ... L:K(x,xL) 

That is, in the training file the first column must be the "ID" of
xi. In testing, ? can be any value.

All kernel values including ZEROs must be explicitly provided.  Any
permutation or random subsets of the training/testing files are also
valid (see examples below).

Note: the format is slightly different from the precomputed kernel
package released in libsvmtools earlier.

Examples:

    Assume the original training data has three four-feature
    instances and testing data has one instance:

    15  1:1 2:1 3:1 4:1
    45      2:3     4:3
    25          3:1

    15  1:1     3:1

    If the linear kernel is used, we have the following new
    training/testing sets:

    15  0:1 1:4 2:6  3:1
    45  0:2 1:6 2:18 3:0 
    25  0:3 1:1 2:0  3:1

    15  0:? 1:2 2:0  3:1

    ? can be any value.

    Any subset of the above training file is also valid. For example,

    25  0:3 1:1 2:0  3:1
    45  0:2 1:6 2:18 3:0 

    implies that the kernel matrix is

        [K(2,2) K(2,3)] = [18 0]
        [K(3,2) K(3,3)] = [0  1]
share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.