Converting categorical variables
Because sklearn requires numerical features as inputs for models, it is important to encode categorical variables into numerical ones. The most common technique, called "one-hot encoding", is straightforward but has high memory consumption. To that end, you will use the technique of hashing, which maps categorical inputs into numerical ones, for each categorical column.
The pandas module is available as pd in your workspace and the sample DataFrame is loaded as df.
This exercise is part of the course
Predicting CTR with Machine Learning in Python
Exercise instructions
- Select the categorical columns by filtering for data type.
- Apply a hash function over each of the categorical columns.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# Get categorical columns
categorical_cols = df.____(
include = [____]).columns.tolist()
print("Categorical columns: ")
print(categorical_cols)
# Iterate over categorical columns and apply hash function
for col in ____:
df[col] = df[col].____(lambda x: ____(x))
# Print examples of new output
print(df.head(5))