Practical Applications of Data Mining

32 Chapter 2 assoCiation rules

In the first part, L

k−1

, the large itemsets generated in the previous pass are

used to generate C

k

, the large candidate k-itemsets. The apriori-gen( ) func-

tion is used for this purpose. In the second part, each itemset in C

k

is counted

in all transactions in the database. Only those itemsets with support that is at

least minsup are collected to generate L

k

.

2.2.2 Apriori-gen( ) Function 

The apriori-gen( ) function takes an argument L

k−1

to generate C

k

, the set

of potentially large k-itemsets. In other words, C

k

is a superset of the set of

all large k-itemsets. The function consists of two steps. First, join operations

are performed to join L

k−1

with L

k−1

. In the next step, we perform pruning

to delete ...