Summary

In this chapter, we were introduced to ML pipelines. Through the myriad of exercises that we implemented, we realized that to squeeze out the best performance from our ML models, we must try various permutations and combinations of features, models, and model parameters. Finding the right combination is indeed a time-consuming process. Using an ML pipeline is a technique that automates this process and spares us a lot of manual experimentation.

Within this chapter, we progressively implemented different parts of an ML workflow. We created a processing engine and added dimensionality reduction and modeling to the processing engine. Later on, we did spot-checking with various models and also performed grid search to find the best parameters. In the process, we identified how ML pipelines made all these processes simpler.

The objective of this chapter was to enable you to carry out different experiments on your ML workflow using a very powerful toolset. It is left to the...