Some nice ML-libraries
I recently had a go at the Kaggle Acquire Valued Shoppers Challenge. This competition was a bit special in that the dataset was 22 GB, one of the biggest datasets they’ve had in a competition. While 22 GB may not quite qualify as big data, it’s certainly something that your average laptop will choke on when using standard methods. Ordinarily I’d reach for scikit-learn for these tasks, but in this case some of the methods in scikit-learn were a bit slow, and some other libraries had nice methods that scikit-learn didn’t have, so I decided to try out some other libraries. In this post I’ll give a brief look at some of the alternative machine learning libraries that I used during the competition.