Q1:

Data objects may belong to more than one class at a time. In such cases it is difficult to assess classification accuracy. Mention your comment on what criteria you would use to compare different classifiers modeled using the same data.

Q2:

a) Classify the following classification techniques to either eager or lazy classification.

Decision tree, case-based reasoning, Bayesian, neural network, k-mean, k-nearest neighbor.

b) Compare and contrast Eager and Lazy classification methods.

Q3:

a) Explain dendrogram. What could be the possible reason(s) for producing two different dendrograms using agglomerative clustering algorithm for the same dataset?

b) In which cases K-Means clustering algorithm fails to give good results?

Q4:

a) What are the advantages of DBSCAN clustering algorithm?

b) Assume, you want to cluster observations into 3 clusters using K-Means clustering algorithm. After first iteration three clusters (C1, C2, C3) have the following observations:

C1: {(4,4), (5,5), (6,6)}

C2: {(0,6), (4,6)}

C3: {(3,9), (11,11)}

Find the cluster centroid of each cluster?