add association learning for the ones with same words (words around it), and the ones without same words (predict to be not so relevant, learn in the same way, but with hidden as~?(using the topic in this session))
whether the word (like some #$#34 title when redirecting) is meaningful -> if there's co-occurences in the titles that have this word
<-- no word can live without other words at all
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
For those keywords that have no occurrence before, there will be no subnodes when querying such keywords. Then it should recognize former patterns that's common with the keywords, like "how to xxx", which are very likely to be consist of stopwords.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
whether the word (like some #$#34 title when redirecting) is meaningful -> if there's co-occurences in the titles that have this word
<-- no word can live without other words at all
For those keywords that have no occurrence before, there will be no subnodes when querying such keywords. Then it should recognize former patterns that's common with the keywords, like "how to xxx", which are very likely to be consist of stopwords.