Practical Applications of Data Mining

2.2 Mining of assoCiation rules in Market Basket Data 33

to be large and is the only member of L

3

. When we generate C

4

with L

3

, it

turns out to be empty, and we terminate.

2.2.4 AprioriTid Algorithm

The AprioriTid algorithm is very similar to the Apriori algorithm in that

both algorithms use the apriori-gen( ) function. In the Apriori algorithm,

the database D is used to count support on every pass when C

k

is generated.

In AprioriTid, however, the database D is not used after the first pass. Rather,

the set C

k

´ is used for counting support. Each member of the set C

k

´

is of the

form <TID, {X

k

}>, where each X

k

is a potentially large k-itemset ...