In this chapter, we went deep down the rabbit hole and looked at several common feature engineering techniques, focusing on both feature selection and feature extraction. We successfully formatted, cleaned, and transformed data so that it could be understood by common machine learning algorithms. We learned about the curse of dimensionality and dabbled a bit in dimensionality reduction by implementing PCA in OpenCV. Finally, we took a short tour of common feature extraction techniques that OpenCV provides for image data.

With these skills under our belt, we are now ready to take on any data, be it numerical, categorical, text, or image data. We know exactly what to do when we encounter missing data, and we know how to transfer our data to make it fit our preferred machine learning algorithm.

In the next chapter, we will take the next step and talk about a specific use case, which is how to use our newly acquired knowledge to make medical diagnoses using decision trees.