I was looking for Teapot like in Julia. Here is Teapot in Python:
Always tune the hyperparameters for our models Always try out many different models Always explore numerous feature representations for our data
We must also consider the following:
There are thousands of possible hyperparameter configurations for every model There are dozens of popular machine learning models There are dozens of popular feature preprocessing methods
This is why it can be so tedious to design effective machine learning systems. This is also why my collaborators and I created TPOT, an open source Python tool that intelligently automates the entire process.