News for hackers. Stuff that matters.

A comment on Top ten algorithms in data mining (2007) [pdf]. Parent comment

Re:Top ten algorithms in data mining (2007) [pdf] by mturmon on Tuesday December 18, 2012 @01:24PM (#4938754)
Crudely: combine the speed of classification trees with the generally low error rates (flexibility) of the SVM.
[ Parent | on HN ]
Re:Top ten algorithms in data mining (2007) [pdf] by sgt101 on Tuesday December 18, 2012 @04:46PM (#4939771)
There is a good implementation in Mahout and it scales well on Hadoop, so its a pretty good out of the box whole data set learning algorithm.

No advantage over taking a decent statistically valid subset and running C4.5 (or a variant with floating point splits) apart from it's less work.

[ Parent | on HN ]