Feature Engineering for Existing Features.png

Feature engineering is a crucial stage in any machine learning project. It allows you to use data to define features that enable machine learning algorithms to work properly.

In this course, you will learn the techniques that will help you create new features from existing features. You’ll start by diving into label encoding which is crucial for converting categorical features into numerical. You’ll also learn about other various types of encoding such as: one-hot, count, and mean, all of which are important for feature engineering. 

In the remaining chapters, you’ll learn about feature interaction and datetime features. In all, this course will show you the many different ways you can create features from existing ones.

Feature Engineering for Machine Learning

# What is feature interaction?

Feature interaction is a method for new features by interacting with two or more existing features. In short, if you have two features of category types, you can create a new feature by joining them together, as shown in the image below.


>**Notice**: Similar operation also works on numerical features, but in different ways.

# The advantages of feature interaction

- The feature space has been expanded, which increases the predictive power of the model. If feature column `A` has 20 values and feature column `B` has 30 values, then the number of `A` and `B` would go up to 600(20*30).
- An increase in the number of features means that the sample distribution is more finely delineated.
- Feature interaction introduces nonlinearity, which also increases the predictive power of the model.

# Some practical tips

- Second-order interaction is most commonly used.
- Higher-order is also used in practice, but it depends on your task and data. Too much higher-order interaction adds too many features, which makes learning harder and can even cause failure.
- Don’t interact with every feature as it is unnecessary. It requires a deep understanding of the domain knowledge of the task.

Ok, let's code now.

# Feature interaction in `pandas`

It's easy to interact with two features (columns) in `pandas`. Just concatenate two columns like below. If you have more than 2 columns, follow the same approach.

```python
df["fea1_fea2"] = df["fea1"].astype('str')+"_"+df["fea2"].astype('str')
```
>**Notice**: If you are not familiar with `pandas`, you can check the course [Data Analysis & Processing with Pandas](https://www.educative.io/courses/data-analysis-processing-with-pandas).

Learn how to add new features by feature interaction.

__default

Feature Interaction

What is feature interaction?