Data Science and Machine Learning (Theory and Projects) A to Z - Feature Engineering: Categorical Features Python

The video tutorial discusses categorical features in data science, using an example from Jack's vendor Passbook Data Science Handbook in Python. It explains how to create a dataset with features like price, rooms, and neighborhood, and demonstrates the use of sklearn's dictionary vectorizer for data vectorization. The tutorial covers one-hot encoding, highlighting its impact on feature expansion and the challenges of high dimensionality. It concludes with a discussion on handling sparse matrices and a preview of the next video on text features.

What are the three features mentioned in the example dataset?

• Value, Rooms, District

• Price, Size, Location

• Price, Rooms, Neighborhood

• Cost, Bedrooms, Area

Which library is used for feature extraction in the video?

• sklearn

• numpy

• matplotlib

• pandas

What is the purpose of one-hot encoding?

• To normalize the data

• To increase the number of samples

• To convert categorical data into numerical format

• To reduce the number of features

How many new features are created from the neighborhood feature using one-hot encoding?

• Three

• Two

• Five

• Four

What is a potential downside of one-hot encoding with high cardinality features?

• It makes the data harder to interpret

• It increases the complexity of the model

• It reduces the accuracy of the model

• It drastically increases the dimensionality of the data

What is a sparse matrix?

• A matrix with mostly zero values

• A matrix with equal number of zero and non-zero values

• A matrix with no zero values

• A matrix with mostly non-zero values

Why might you set the sparse parameter to true?

• To decrease the number of samples

• To make the matrix more memory efficient

• To increase the number of features

• To improve the accuracy of the model