work on two data mining exercises using Orange: one on association rules and the other on time-series analysis. Take screenshots of your work in Orange wherever specified (show all the relevant portions of the screen), and copy & paste them next to your answers. Please make sure to number your answers.
A. Association Rules
In this exercise, you will be using the Foodmart data to identify frequent itemsets and generate association rules based on the shopping data. Launch Orange and load the foodmart.pkl file.
1. Use the Frequent Itemsets widget to identify the items that appear frequently appear together in a shopping cart. For finding the itemsets, use a minimal support of 2%.
a. Which is the item that is most frequently purchased? In what percent of the transactions does it appear? (take screenshot)
b. Which store sells this item the most?
c. Which are the three items that are bought most frequently along with the item you identified in 1a?
d. Changes the minimal support to 0.5% and the minimum items to 3. Based on these new parameter values, identify the item that is bought most frequently and the items that appear on the shopping cart along with it. (take screenshot)
2. Use the Associate Rules widget to generate the association rules. For finding the rules, use a minimum support of 0.01% and a minimum confidence of 90%.
a. Sort the rules based on lift and identify the two rules with the highest lift (specify them in order of lift). Explain clearly what these lift value mean. (take screenshot)
b. Filter the rules from 2a by selecting only those that include Wine in the antecedent. Among these rules, which one has the highest lift? Explain clearly what this lift means. (take screenshot)
