With recent developments in big data, we’ve been given more access to data in general and high-dimensional data. Consequently, the performance of machine learning models has improved by a large margin. On the other hand, there are significant features often collected or generated by different sensors and methods that can influence the model accuracy in a harmful way that needs careful consideration, Not only that, but these features can demand a lot of computational resources to build and maintain the model.

For that, we need handy processes that contribute to the machine learning pipeline to build great models even with these kinds of features, In this book about feature engineering and feature selection techniques for machine learning with python in a hands-on approach, we will explore pretty much all you’ll need to know about feature engineering and feature selection.

Specifically, we’ll learn how to modify dataset variables to extract meaningful information to capture as much insight as possible, filter out unneeded features leaving datasets and their variables ready to be used in machine learning algorithms.

What's inside?

This book is divided into two parts— Feature Engineering and Feature Selection, we will start with

feature engineering first, then we will move to the other section of feature selection.

Here are brief descriptions of each of the sections:

Part I: Feature Engineering

• Feature Types: Or variables types—we’ll learn about continuous, discrete, and categorical variables (which can be nominal or ordinal), alongside time-date and mixed variables.
• Common Issues: This chapter will discuss different issues you’ll see in real-world datasets like missing data, variable distribution, data imputations, outliers, and others.
• Dealing with Missing Values: We’ll examine the major techniques used to fill the missing data in your dataset.
• Categorical Encoding: This chapter will discuss the different techniques to transform categorical variables into numbers—frequency encoding, one-hot encoding, and more.
• Feature Transformation: We’ll explore the mathematical transformations you can apply to alter the distribution of numerical variables, like logarithmic or reciprocal transformations.
• Variable Discretization: This chapter looks at the procedures to discretize variables, like equal width, equal-frequency, discrete decision using decision trees, clustering, and more.
• Dealing with Outliers: This chapter will show how to identify outliers and remove them from your dataset.
• Feature Scaling: We’ll cover several techniques to scale features, like standardization, scaling to the minimum and maximum, scaling to the unit length of the vector, and more.
• Handling Time-Date and Mixed Variables: We’ll discuss various ways to create new features from date, time, and mixed variables.
• Engineering Geospatial Features: We’ll see how to handle geospatial features that are (most of the time) represented as longitude and latitude.
• Resampling Imbalanced Datasets: We will learn how to apply a technique called resampling that can solve an issue where the classes in a given dataset are not represented equally.
• Advanced Feature Engineering: We’ll look at advanced categorical encoding, advanced outlier detection, automated feature engineering, and more.

Part II: Feature Selection

• Filter Methods: Rely on the features’ characteristics without using any machine learning algorithm. Very well-suited for a quick “screen and removal” of irrelevant features.
• Wrapper methods: Consider the selection of a set of features as a search problem, then uses a predictive machine learning algorithm to select the best feature subset. In essence, these methods train a new model on each feature subset, which makes it very computationally expensive. However, they provide the best performing feature subset for a given machine learning algorithm.
• Embedded methods: Just like the wrapper methods, embedded methods take the interaction of features and models into consideration. They also perform feature selection as part of the model construction process, and they are less computationally expensive.
• Advanced Feature Selection: We’ll see more advanced techniques that use deep learning, heuristic searches, and several other ways to select features.

In each chapter of this book, we’ll practically learn various techniques and examine their advantages, limitations, the assumptions that each method makes, and when you should consider applying each technique.