oh my god!!!!!!!!!!!!! They are highlighting the answer. That's so smart. You know how they do that? One way is to calculate the mutual information between two terms, for example they see if these terms always co-occur, so it means that they have high mutual information.
This paper was published in SIGIR 2009, it explains how it can be done. "Learning to Rank for Quantity Consensus Queries" BTW, it seems that blogspot doesn't let you put link in your comment. That suxxx!!!!!!!!!!
oh my god!!!!!!!!!!!!! They are highlighting the answer. That's so smart. You know how they do that? One way is to calculate the mutual information between two terms, for example they see if these terms always co-occur, so it means that they have high mutual information.
ReplyDeleteThis paper was published in SIGIR 2009, it explains how it can be done.
ReplyDelete"Learning to Rank for Quantity Consensus Queries"
BTW, it seems that blogspot doesn't let you put link in your comment. That suxxx!!!!!!!!!!