> s1 <- data.frame(x=c(1,2,3), y=c(7,8,7),series=rep("s1", 3)) > s2 <- data.frame(x=c(1,2,3,4), y=c(6,7,6,7), series=rep("s2", 4)) > s <- rbind(s1, s2) > qplot(x, y, data=s, geom=c('point', 'line'), color=series)
You should use something like tf-idf to normalize your cooccurance counts for your recommender, otherwise you'll just end up recommending the most globally popular products.
It's called "pic for pic" and I hear the kids are really into it.
Looks like the PMML issue has just been re-opened on scikit-learn's github: https://github.com/scikit-learn/scikit-learn/issues/1596
Using a mutable "last active time" from some sort of users table is pretty dangerous for post-hoc analysis because if (when!) something goes wrong, the information is overwritten and gone forever. It's better to use…
> s1 <- data.frame(x=c(1,2,3), y=c(7,8,7),series=rep("s1", 3)) > s2 <- data.frame(x=c(1,2,3,4), y=c(6,7,6,7), series=rep("s2", 4)) > s <- rbind(s1, s2) > qplot(x, y, data=s, geom=c('point', 'line'), color=series)
You should use something like tf-idf to normalize your cooccurance counts for your recommender, otherwise you'll just end up recommending the most globally popular products.
It's called "pic for pic" and I hear the kids are really into it.
Looks like the PMML issue has just been re-opened on scikit-learn's github: https://github.com/scikit-learn/scikit-learn/issues/1596
Using a mutable "last active time" from some sort of users table is pretty dangerous for post-hoc analysis because if (when!) something goes wrong, the information is overwritten and gone forever. It's better to use…