Alternative way of classifying with IForest
Until now, you have been using the .fit_predict() method to fit IForest and generate predictions simultaneously. However, pyod documentation suggests using the fit function first and accessing the inlier/outlier labels_ via a handy attribute.
You will practice this on the big_mart dataset.
Este ejercicio forma parte del curso
Anomaly Detection in Python
Instrucciones del ejercicio
- Fit (only
fit) theIForest()estimator tobig_mart. - Access the training labels and store them as
labels. - Use
pandassubsetting onbig_martto filter the outliers intooutliers.
Ejercicio interactivo práctico
Prueba este ejercicio completando el código de muestra.
iforest = IForest(n_estimators=200)
# Fit (only fit) it to the Big Mart sales
____
# Access the labels_ for the data
labels = iforest.____
# Filter outliers from big_mart
outliers = ____[____]
print(len(outliers))