Exercise

LOF calculation

kNN is useful for finding global anomalies, but is less able to surface local outliers. In this exercise, you'll practise using the lof() function to calculate local outlier factors for the wine data.

lof() has the arguments:

x: the data for scoring
k: the number of neighbors used to calculate the LOF

Instructions

100 XP

Calculate the local outlier factor using the 5-nearest neighbors.
Append the LOF to the wine data set as a new column called score.

Take Hint (-30 XP)