kNN distance score
Once the kNN distance matrix is available, the nearest neighbor distance score can be calculated by averaging the nearest neighbor distances for each point.
Large values of the distance score can be interpreted as indicating the presence of unusual or anomalous points.
This exercise is part of the course
Introduction to Anomaly Detection in R
Exercise instructions
- Calculate the k-nearest neighbor distance score using the
rowMeans()
function. - Print the row index of the point with the largest distance score.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# Calculate the 5 nearest neighbors distance
wine_nn <- get.knn(wine, k = 5)
# Create score by averaging distances
wine_nnd <- rowMeans(___)
# Print row index of the most anomalous point