Subject: Machine Learning

Question Number 2 :

1. Suppose we have market basket data consisting of 100 transactions and 20 items. Compute the total number of possible association rules. If the support for item A is 25%, the support for item B is 90% and the support for itemset {A,B} is 20%. Let the support and condence thresholds be 10% and 60%, respectively. Compute the condence of the association rule {A} {B} & {B} {A}.
2. Consider the following data of software repository which represents set of source files changed together in six different transactions (TID). Transaction is an event in which developers change one or more source files together to incorporate changes in software.

Table-1

TID	Items (source files change together)
1	File1, File2, File3
2	File1, File2
3	File1, File4, File5
4	File5, File4
5	File5, File3
6	File1, File4, File5

Use Table-1 data and apply APRIORI algorithm to find the most frequent association rules which can be used to predict those source files which are most frequently change together. The minimum support and confidence values are 33.34% and 60.0% respectively.