Interesting approach. It seems a shame not to mix their expert knowledge/recommendations with the collaborative filtering data that they collect when people use it. I think the best recommendation systems will try to leverage all these types of data. I also think it's neat to supply these kind of recommendations in web-search format: as a ranked list, given a query/set of queries. A project that tries to do that for movies, music and books is here here. After all recommending web pages, or recommending music, can't some parallels be drawn here? Come on Google, fix it for us!
Interesting approach. It seems a shame not to mix their expert knowledge/recommendations with the collaborative filtering data that they collect when people use it. I think the best recommendation systems will try to leverage all these types of data.
I also think it's neat to supply these kind of recommendations in web-search format: as a ranked list, given a query/set of queries. A project that tries to do that for movies, music and books is here.
After all recommending web pages, or recommending music, can't some parallels be drawn here? Come on Google, fix it for us!
Slashdot Mirror
Interesting approach. It seems a shame not to mix their expert knowledge/recommendations with the collaborative filtering data that they collect when people use it. I think the best recommendation systems will try to leverage all these types of data. I also think it's neat to supply these kind of recommendations in web-search format: as a ranked list, given a query/set of queries. A project that tries to do that for movies, music and books is here here. After all recommending web pages, or recommending music, can't some parallels be drawn here? Come on Google, fix it for us!
Interesting approach. It seems a shame not to mix their expert knowledge/recommendations with the collaborative filtering data that they collect when people use it. I think the best recommendation systems will try to leverage all these types of data. I also think it's neat to supply these kind of recommendations in web-search format: as a ranked list, given a query/set of queries. A project that tries to do that for movies, music and books is here. After all recommending web pages, or recommending music, can't some parallels be drawn here? Come on Google, fix it for us!