As a concluding topic, we will discuss how to wrap together the operations of transformation and selection we have seen so far, into a single command, a pipeline that will take your data from source to your machine learning algorithm.
Wrapping all your data operations into a single command offers some advantages:
Your code becomes clear and more logically constructed because pipelines force you to rely on functions for your operations (each step a function)
You treat the test data in the same exact way as your train data without code repetitions or possibility of any mistake in the process
You can easily grid-search the best parameters on all the data pipelines you devised, not just on the machine learning hyperparameters
We distinguish between two kinds of wrappers depending on the data flow you need to build: serial or parallel.
Serial processing means that your transformation steps are dependent one on the other, and consequently they have to be executed in...