Anomaly score
Your visualization suggested that thyroid disease could be detected from anomalous hormone measurements.
In this exercise you'll use an isolation forest to generate an anomaly score for thyroid levels, and compare the resulting score against the true disease status.
This exercise is part of the course
Introduction to Anomaly Detection in R
Exercise instructions
- Fit an isolation forest to the
thyroid
hormone measurements. - Generate anomaly scores for the thyroid data and append the result to
thyroid
as the new columniso_score
. - Use the
boxplot()
function to compare the score distribution for patients with and without thyroid disease, using thelabel
column.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# Fit isolation forest
thyroid_forest <- iForest(___, ___ = 200, phi = 100)
# Anomaly score
thyroid$iso_score <- predict(thyroid_forest, ___)
# Boxplot of the anomaly score against labels
boxplot(___ ~ ___, ___, col = "olivedrab4")