Optimizing Models and Using AutoML

So far, we've looked at a few machine learning (ML) models for classification and regression: simple linear models (linear regression and logistic regression), k-nearest neighbors (KNN), and Naïve Bayes for classification. As we will see in these next few chapters, there are other models that are commonly used in ML and data science. This chapter will cover how to choose between models and how to optimize models. Specifically, we'll cover:

• Hyperparameter optimization with random, grid, and Bayesian searches
• Using learning curves to optimize the amount of data needed and diagnose ML models
• Optimizing the number of features with recursive feature selection
• Using the pycaret AutoML Python package

Let's get started with hyperparameter optimization using a few different search methods.