Letter #9 - In which we go a little further into the water, but manage not to drown.
“Always go a little further into the water than you feel you’re capable of being in.”
But not in actual water ‘cause that’s scary, that’s what we have metaphors for. Read the whole thing, it’s good. Then maybe also read Seth Godin’s Don’t expect applause.
As you stand
Under the bar lights
And the band plays some song
About forgetting yourself for a while
A visceral song they say, and I tend to agree. The version they recorded live from the Walt Disney Concert Hall is also beautiful.
Ensemble methods are meta-algorithms that combine several machine learning techniques into one predictive model in order to decrease variance (bagging), bias (boosting), or improve predictions (stacking).
A nice overview of the three main types of ensembles, including algorithms, how they work, and visualizations of their respective performance. As Kaggle’s #1 contender, Bestfitting recently mentioned, you do need ensembles for achieving top performance, but you can’t count on them doing magic with only weak-performing models to work with.
Not sure what you just read? Take a look at this post.
Enjoyed it? Subscribe to my tiny newsletter and get new issues straight to your inbox. Every week, on Tuesdays.