From: fancyerii@gmail.com   
      
   Hi, all.   
       I'd like to use SVM to finish a POS tagging task. And I have   
   chosen some features such as current word, privous word, words   
   trigram(w-2,w-1,w0).   
      Because SVM requires a vector to describe a sample. One solution I   
   can come up with is that Collecting all the words occured in the   
   training set as a dictionary. And mapping each word into a integer   
   value. But what about the words which have not appeared in training   
   set? Mapping all these words into a specific integer such as Infinite?   
      In the svmlight website, there is a toolkit named Tree Kernel. Is   
   it helpful to mapping a string into a vector?   
     Thanks!   
      
   [ comp.ai is moderated ... your article may take a while to appear. ]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)