Assume that the K-medoid clustering algorithm has been applied to a data set described by two descriptive features. The table below shows the instances assigned to cluster 1. What is the ID of the instance that should represent the cluster? Assume that Manhattan distance is used.

ID   d1   d2

4     3     4

5     6     2

9     6     4

13   7     3

15   7     4

17   7     6

19   8     5