Throughout the previous chapters, you have seen many different techniques applied to modeling problems and most of those techniques, although they seem simple, include many parameters that ultimately affect the outcome of your efforts. Many modeling problems require AutoML to be represented as a search problem and in the majority of cases, there are only sub-optimal solutions to be found.
In a broader sense, modeling is just a mapping between your input data and output data. As a result, you will be able to infer the output where new input data arrives with unknown output. In order to achieve your objective, you need to think about your experiment design and configure your environment accordingly, since you really don't know what will be the best-performing ML pipeline—but let's stop for a second and step back.
Implementing performant...
