a qhestion about AI::Categorizer::KnowledgeSet

[email protected] Sat, 20 Nov 2004 00:41:22 +0800 (CST)
Newsgroups perl.ai
Message-ID <[email protected]>
Dear all:
     I've got a problem of term weighting of text classification.
     The parameter of ftidf_weighting is used to set the term weighting for training data only.
     But, how do we set term weighting for a testing document? 
     shouldn't we assign testing data and training data the same weighting method?
     For example: decide to set tf*idf for term weighting. we can use 
     
        my $k = new AI::Categorizer::KnowledgeSet(ftidf_weighting => 'tfx' );
 
     for training data.
     However, I do not know hoe to use the same set 'tfx' for a new document or a document collection.   
     Please give me some advice!
     Thanks!



---------------------------------
Yahoo!穢_翹簪Messenger6.0
禮籀禮Y簧礎糧翻穫穠繙禮Y簧糧q簞T瞼@竅癒A瞼禮Y瞻U繡羹糧繚s穠穢癒I簞穡瞻W瞻U繡羹