1. Learn
    2. /
  2. Courses
    3. /
  3. Introduction to Anomaly Detection in R

Exercise

Checking convergence

The anomaly score from an isolation forest usually don't change after a certain number of trees have been grown. This is called convergence, and can be checked by comparing the scores generated by forests with different numbers of trees. If the scores differ greatly, then this might suggest that more trees are required.

In this exercise, the scores for isolation forests with different numbers of trees have been already calculated for you and are contained in the data frame wine_scores.

Instructions 1/2

undefined XP
    1
    2
  • Produce a scatterplot showing the scores for 2000 trees as the y-axis variable and the score for 1000 trees as the x-axis variable.