Polynomial features in Data Analysis Python - Cheat Sheet & Quick Revision

Recall & Review

beginner

What are polynomial features in data science?

Polynomial features are new features created by raising existing features to a power (like squared or cubed) or by multiplying features together. They help models learn more complex patterns.

beginner

Why do we use polynomial features in machine learning?

We use polynomial features to allow simple models, like linear regression, to fit curved or more complex relationships between variables.

intermediate

How does the degree parameter affect polynomial features?

The degree controls the highest power used. For example, degree 2 adds squared terms and interaction terms, while degree 3 adds cubes and more complex combinations.

intermediate

What is an interaction term in polynomial features?

An interaction term is a feature made by multiplying two or more original features together, showing how they work together to affect the outcome.

intermediate

What is a potential downside of using high-degree polynomial features?

Using very high-degree polynomial features can make the model too complex, causing it to fit noise instead of real patterns, which is called overfitting.

What does a polynomial feature of degree 2 include?

A Original features, squared features, and interaction terms

B Only original features

C Only squared features

D Only interaction terms

Degree 2 polynomial features include the original features, their squares, and products of pairs of features (interaction terms).

Why might adding polynomial features improve a linear regression model?

A It reduces the number of features

B It allows the model to capture nonlinear relationships

C It makes the model run faster

D It removes noise from data

Polynomial features let the model learn curved relationships, not just straight lines.

What is an interaction term in polynomial features?

A A feature removed from the dataset

B A feature raised to a power

C A feature normalized between 0 and 1

D A feature created by multiplying two or more original features

Interaction terms are products of features, showing combined effects.

What risk increases when using very high-degree polynomial features?

A Overfitting the model to noise

B Underfitting the data

C Reducing model accuracy

D Losing original features

High-degree polynomials can make the model too complex, fitting noise instead of true patterns.

Which Python library provides a tool to create polynomial features easily?

A pandas

B matplotlib

C scikit-learn

D numpy

scikit-learn has a PolynomialFeatures class to generate polynomial features.

Explain in your own words what polynomial features are and why they are useful.

Think about how simple features can be combined or raised to powers to help models learn curves.

Describe the potential risks of using high-degree polynomial features in a model.

Consider what happens when a model becomes too flexible.