Chapter 7: Feature Engineering
Now that we have covered some considerable ground on how to shape our data as needed, let's talk about feature engineering.
If you want to create a machine learning model, you input data. This input data includes the features that an algorithm needs to create a model. These features need to have specific characteristics; for example, it cannot have null values or the data needs to comply and have specific probability distributions.
With featuring engineering, you can prepare the input dataset so that it complies with the algorithm's requirements, and also improve the performance of the machine learning model, thereby creating new features with data we already have.
So, in this chapter, we will be covering the following topics:
- Handling missing values
- Handling outliers
- Binning
- Variable transformation
- One-hot encoding
- Feature splitting
- Scaling
