Leggere, an internet book retailer, is interested in better understanding the purchase decisions of its customers. For a set of 2,000 customer transactions, it has categorized the individual book purchases comprising those transactions into one or more of the following categories: Novels, Willa Bean series, Cooking Books, Bob Villa Do-It-Yourself, Youth Fantasy, Art Books, Biography, Cooking Books by Mossimo Bottura, Harry Potter series, Florence Art Books, and Titian Art Books. Leggere has conducted association rules analysis on this data set and would like to analyze the output. Based on a minimum support of 200 transactions and a minimum confidence of 50%, the table below shows the top 10 rules with respect to lift ratio.

Antecedent	Consequent	Support for A and C	Confidence	Lift Ratio
BotturaCooking	Cooking	0.227	1.00	1.16
Cooking, BobVilla	Art	0.205	0.54	1.12
Cooking, Art	Biography	0.204	0.61	1.10
Cooking, Biography	Art	0.204	0.53	1.10
Youth Fantasy	Novels, Cooking	0.245	0.55	1.075
Cooking, Art	BobVilla	0.205	0.61	1.055
Cooking, BobVilla	Biography	0.218	0.58	1.04
Biography	Novels, Cooking	0.293	0.53	1.035
Novels, Cooking	Biography	0.293	0.57	1.035
Art	Novels, Cooking	0.249	0.52	1.01

(a)	Explain why the top rule "If customer buys a Bottura cooking book, then they buy a cooking book," is not helpful even though it has the largest lift ratio and 100% confidence.
	A Bottura cooking book is a specific type of cooking book, so whenever a customer purchases a Bottura cooking book they - Select your answer -mustmust notItem 1 also have bought a cooking book. Thus, the confidence of "If a customer purchases a Bottura cooking book, then they purchase a cooking book" is: %. Thus, any rule with the antecedent being a subset of the consequent will have a confidence of % and provides no information. Furthermore, if we would interpret this rule as suggesting that if a customer is interested in a specific type of cooking book that they - Select your answer -maymay notItem 4 be interested in other types of cooking books, this still provides no profound insight.
(b)	Explain how the confidence of 53% and lift ratio of 1.10 was computed for the rule "If a customer buys a cooking book and a biography book, then they buy an art book." Interpret these quantities.
	A confidence of 53% means that for 53% of the transactions when a cooking book and biography are purchased, an art book - Select your answer -isis notItem 5 purchased. A lift ratio 1.10 means that a transaction in which a cooking book and biography is purchased is 10% - Select your answer -morelessItem 6 likely to also have purchased an art book than a randomly-selected transaction.
(c)	Based on these top 10 rules, what general insight can Leggere gain on the purchase habits of these customers?
	In the top 10 rules, there are two rules involving the item set {cooking, Bob Villa, art} and two rules involving the item set {biography, novels, cooking} and {cooking, art, biography}. Leggere - Select your answer -maymay notItem 7 be well served to investigate promotions involving these respective item sets.
(d)	What will be the effect on the rules generated if Leggere decreases the minimum support and reruns the association rules analysis?
	Decreasing the minimum support will generate a - Select your answer -largersmallerItem 8 pool of rules and therefore will likely result in association rules with larger lift ratios. A rule with a - Select your answer -largersmallerItem 9 lift ratio suggests that it is effective at identifying transactions in which the consequent occurs. However, the transactions involving the item set of the association rule may not occur as frequently due to the lowered minimum support. Thus, the association rule - Select your answer -maymay notItem 10 be spurious, or the item set may apply to too few transactions to be a valuable business opportunity.
(e)	What will be the effect on the rules generated if Leggere decreases the minimum confidence and reruns the association rules analysis?
	Decreasing the minimum confidence will generate a - Select your answer -largersmallerItem 11 pool of rules and therefore likely result in association rules with larger lift ratios. A rule with a - Select your answer -largersmallerItem 12 lift ratio suggests that it is effective at identifying transactions in which the consequent occurs. However, if this rule has a - Select your answer -lowhighItem 13 confidence then this means that there also is a low probability that the consequent will occur given the antecedent. That is, as the confidence of an association rule - Select your answer -decreasesincreasesItem 14 , the relationship between the antecedent and consequent will fail more often. Depending on the profit margin associated with the transaction, this may make the association rule not a viable business opportunity.