Prev: table
Next: Random matrix with boundaries
From: Mohammad A. Mezher on 11 Apr 2010 14:29
hi everyone,

svm depends on the kernel function so many literature focusing on trying to change kernel function or their parameters...

lastly, when i was wandering i found the following kernel function
K=(Xi min_s (Xi min_s Xj)) •(Xj min_s (Xj min_s Xi));

knowing that "min_ s" referring to scalar minus operator. pls notice that there are two X (Xi, Xj).

This new kernel function found by Yue Jiang et. al in the paper "Automatic SVM Kernel Function Construction Based on Gene Expre ssion Programming" for breast dataset.

the confusing thing is when i applied this kernel i got a scalar value not a m*m matrix.

am i missing something here???
 | 
Pages: 1
Prev: table
Next: Random matrix with boundaries