Question

A large supermarket tracks sales data by stock- keeping unit (SKU) for each item, such as “pasta”, “rice”, “egg”, “soup” is identified by a numerical SKU. The supermarket has a database of transactions where each transaction is a set of SKUs that were bought together. The database of transactions consist of following itemsets:

Transaction ID	Itemsets
T1	egg, soup
T2	rice, egg, soup
T3	rice, soup
T4	pasta, rice, egg, soup
T5	rice, egg
T6	pasta, rice, soup
T7	pasta, rice

Solve the above problem using Apriori Algorithm (Association Rule Mining) with minimum support of 3 and write each and every step.

Answer 1

Answer : Given that,

Transaction ID T1 T2 T3 T4 T5 T6 T7 Itemsets egg, soup rice, egg, soup rice, soup pasta, rice, egg, soup rice, egg pasta, rice, soup pasta, rice

Minimum support count = 3

Step! Itemset Support count 4 egg soup rice pasta 5 6 3

All the items have support count more than 3.

Support count egg.soup Wlw Legg, rice egg, rice soup, rice legg-pasta Soup, vice soup, pasta Step2 : Generate candidate set of size 2. Itemset support count Itemset 898 soup 3 3 eliminate items with support count <3 4 4 2 X 3 rice, pastal 3 Step3 - Generate candidate set of size 3 Itenset support Count legg soup, nice diminate all items, since all have soup, rice, pasta 2 X support count <3 egg, rice, pasta 1 frice, pasta 2

Hence, the frequent itemsets are egg-soup, egg-rice, soup-rice, rice-pasta.

Next , we generate association rules for each frequent itemset by calculating its confidence percent using the given formulae :

For 3 4 = 0.75- 75% olw fla - 0.60 = 60% Foo egg, sice = 3 4 Confidence (A>B) = Support count (AUB) Support Colont (A) egg, soup => confidence (egg → soup) = confidence (soup> ogg) = } confidence (eggərice) -0.75 -75% confidence (rice- egg) = 3 50% confidence (soup->nice) - 4 confidence (rice-1 soup) = 4 For rice, pasta con afidence (rice-pasta) 50% Confidence (pasta rice) = 3 = 0.50 For soup, rice 0.80 = 80% 5 = 0.66 = 66% 6 3 6 SO. SO = 100% 3