Exam DISC-DA Practice Question 226
Study guide: Exam DISC-DA
A homeowners analyst runs K-means clustering on policy characteristics (dwelling age, square footage, construction, distance to coast and protection class), with no loss data, and gets five well-separated clusters. She proposes using the five clusters as new rating segments. What should be done before the clusters are used as segments?
- A) Test whether the clusters differ in loss experience and are credible and stable
- B) Nothing further, since K-means found five clearly separated clusters
- C) Rerun K-means with more clusters until every cluster has the same loss ratio
- D) Replace the clusters with alphabetical groupings of the protection classes
Answer this question and see the worked solution