LOF with outlier probabilities
As always, double-check that the chosen contamination level is trustworthy by filtering the outliers with a probability threshold. The syntax is the same as with KNN
.
LOF
estimator has already been imported, and the females_transformed
dataset is also available.
This exercise is part of the course
Anomaly Detection in Python
Exercise instructions
- Instantiate
LOF()
with 20 neighbors. - Calculate outlier probabilities into
probs
. - Create a boolean mask named
is_outlier
that returns true values where the outlier probability is over 50%. - Use
is_outlier
to filter the outliers fromfemales_transformed
.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# Instantiate an LOF with 20 neighbors and fit to the data
lof = ____
lof.____
# Calculate probabilities
probs = ____
# Create a boolean mask
is_outlier = ____
# Use the boolean mask to filter the outliers
outliers = ____
print(len(outliers))