Non-aggressive methods of dealing with outliers
Trimming a distribution is one of the easiest ways of dealing with an outlier without dropping it. If you know the natural range of values a distribution must take related to your business problem, you can use that knowledge to trim the distribution.
Or, if you have some previous knowledge about which specific values are considered outliers based on domain knowledge or experience, you can also replace them with hard-coded values.
This exercise is a chance to practice these two techniques on the apple stocks dataset.
Cet exercice fait partie du cours
Anomaly Detection in Python
Exercice interactif pratique
Essayez cet exercice en complétant cet exemple de code.
# Find the first percentile of Volume
percentile_first = ____