Exam DISC-DA Practice Question 234
Study guide: Exam DISC-DA
An analyst describes small business policies by 40 yes/no indicators, one for each optional endorsement; any one policy carries only two or three endorsements. She wants to cluster policies whose endorsement choices are alike. Why is the Jaccard dissimilarity a better choice here than simple matching (the share of the 40 indicators on which two policies disagree)?
- A) Jaccard leaves out the endorsements that neither of the two policies carries
- B) Jaccard gives more weight to the endorsements with the highest premiums on each policy
- C) Jaccard standardizes each indicator to mean zero and standard deviation one first
- D) Jaccard treats the indicators as numeric, so differences can be squared and summed
Answer this question and see the worked solution