1. PRISM by hand: With the weather dataset (10 instances) shown in the notes for this week, follow the PRISM procedure to construct rules for Play = Yes. Show all your work, and clearly present the process you follow. If there are any ties, state that you are breaking them arbitrarily before choosing a rule to continue. Be sure to write down your final rule set.

2. The notes explain that the 2-itemset (X,Y) can construct two rules: 1) If X then Y and 2) If Y then X for the items X and Y. Must both rules have the same confidence? Must both rules have the same support? Demonstrate your answer with examples or counterexamples from the nominal weather dataset in the notes.

3. Using the 10-instance weather dataset from the notes for this week, what is the support of the rule "If temperature = mild then Windy = true"? What is the confidence of the rule? Is this an association rule or a classification rule? Justify your answer by showing work or explaining reasoning.

4. Using the 10-instance weather dataset from the notes for this week, consider the 3-itemset (outlook = sunny, temperature = hot, humidity = high). What rules with a 1-itemset as consequent can be created from this itemset? Calculate the confidence of each rule using the formula of slide/page 22 showing your work clearly. That forumula is given there as p = t(i1,i2,,in)/t(i1,i2,ij-1,ij+1,,in).