Question 25 (1 point)

In the original k-means clustering algorithm, the parameter k is

Question 25 options:

	determined solely by the user
	initially set by the user, but then the algorithm converges on a value for k
	determined solely by the algorithm

(The following two questions are based on this description)

Suppose we are clustering the following set of instances using k-means clustering, with k = 2: {(1,1), (1,2), (1,3), (1,4), (3,1), (4,1), (5,1)}. Assume that the initial centers are C1 at (1,4) and C2 at (3,1). Note, the pairs show values for (x, y), where x and y are the attributes for an instance.

Question 26 (1 point)

Which of the following shows the initial clusters?

Question 26 options:

	Cluster 1: {(1,3), (1,4)}, Cluster 2: {(1,1), (1,2), (3,1), (4,1), (5,1)}
	Cluster 1: {(1,1), (1,2), (1,3), (1,4)}, Cluster 2: {(3,1), (4,1), (5,1)}
	Cluster 1: {(1,2), (1,3), (1,4)}, Cluster 2: {(1,1), (3,1), (4,1), (5,1)}
	Cluster 1: {(1,1), (1,2), (1,3), (1,4), (3,1)}, Cluster 2: {(4,1), (5,1)}

Question 27 (1 point)

If we continue the clustering process until convergence, which of the following shows the final clusters?

Question 27 options:

	Cluster 1: {(1,3), (1,4)}, Cluster 2: {(1,1), (1,2), (3,1), (4,1), (5,1)}
	Cluster 1: {(1,1), (1,2), (1,3), (1,4)}, Cluster 2: {(3,1), (4,1), (5,1)}
	Cluster 1: {(1,1), (1,2), (1,3), (1,4), (3,1)}, Cluster 2: {(4,1), (5,1)}
	Cluster 1: {(1,2), (1,3), (1,4)}, Cluster 2: {(1,1), (3,1), (4,1), (5,1)}

Question 28 (1 point)

Consider the incremental clustering algorithm, at a certain step, we have formed a tree as shown below, when a new instance f comes in, where might we insert this instance? Suppose that among the root and the five leaf nodes (a-e), a has the highest category utility (as a host) and b is the runner-up.

Question 28 options:

	as a new leaf of the root
	we create a new internal node with a and f as leaf nodes
	we create a new internal node with a, b and f as leaf nodes
	b, c
	a, b, c

Question 29 (1 point)

In hierarchical agglomerative clustering algorithm, the similarity (or distance) between two clusters can be decided by

Question 29 options:

	the similarity (or distance) between the two closest members of the clusters
	the similarity (or distance) between the two farthest members of the clusters
	the similarity (or distance) between the centroids of the two clusters
	average of the similarity (or distance) between all members of the clusters
	all of the above

Question 30 (1 point)

In a multi-instance learning problem (where a single example is a bag of instances), the training set would have

Question 30 options:

	a class associated with each bag
	a class associated with each instance in each bag (so a bag might have several classes in it)
	no class associated with either bags or instances