simplestringkernel.m

来自「Proteins _ String Kernel」· M 代码 · 共 28 行

%Computes the string kernel for contiguous strings of length seqLength
%between the strings in proteinStrings.

%Arthur Gretton

function [K]=simpleStringKernel(seqLength,proteinStrings);

disp('Computing spectrum kernel...')

totalStringNum =length(proteinStrings);
sequenceFeatures = zeros(totalStringNum,26^seqLength);  %rows represent different data points

for l=1:totalStringNum
  singleString = char(proteinStrings(l));
  for m=1:length(singleString)-seqLength+1  
    substring=singleString(m:m+seqLength-1);
     featureIndex=1;
     for n=1:seqLength
       featureIndex = featureIndex+(substring(n)-'a')*26^(n-1);
     end     
     sequenceFeatures(l,featureIndex)=sequenceFeatures(l,featureIndex)+1;
  end
  l;
end

K=sequenceFeatures*sequenceFeatures';