Part A: (2.5 Marks)
- Use the training data in Table 1 to construct a confusion matrix for the decision tree shown in Figure 1 (Class Label = Life Insurance Promotion).
- Write the production rules for the decision tree shown in Figure 2.
- Can you simplify any of your rules without compromising rule accuracy (tree pruning)?
Table 1
Figure 1
Figure 2
Part B: (2.5 Marks)
Given the training set D in Table 1, apply the decision tree induction algorithm for classification to construct the decision tree to predict the class label (Life Insurance Promotion (yes/no)). Explain the steps you have taken to construct the decision tree in detail.
Part C (2.5 Marks)
Describe each of the following clustering algorithms in terms of the following criteria: (1) shapes of cluster that can be determined; (2) input parameters that must be specified; and (3) limitations and advantages.
- K-means
- K-medoids
- Agglomerative Clustering
- DBSCAN
Sample Solution