Features with low variance

In the previous exercise you established that 0.001 is a good threshold to filter out low variance features in head_df after normalization. Now use the VarianceThreshold feature selector to remove these features.

Diese Übung ist Teil des Kurses

Dimensionality Reduction in Python

Kurs anzeigen

Anleitung zur Übung

Create the variance threshold selector with a threshold of 0.001.
Normalize the head_df DataFrame by dividing it by its mean values and fit the selector.
Create a boolean mask from the selector using .get_support().
Create a reduced DataFrame by passing the mask to the .loc[] method.

Interaktive Übung

Vervollständige den Beispielcode, um diese Übung erfolgreich abzuschließen.

from sklearn.feature_selection import VarianceThreshold

# Create a VarianceThreshold feature selector
sel = ____(threshold=____)

# Fit the selector to normalized head_df
sel.fit(____ / ____)

# Create a boolean mask
mask = sel.____

# Apply the mask to create a reduced DataFrame
reduced_df = head_df.loc[____, ____]

print(f"Dimensionality reduced from {head_df.shape[1]} to {reduced_df.shape[1]}.")

Code bearbeiten und ausführen