Mining frequent patterns and association rules using similarities
Expert Systems with Applications: An International Journal
| Hi-index | 0.00 |
The TSAA(transitive support association Apriori) algorithm of this paper discovers transitively associated relationships as well as the frequent itemsets. It utilizes the join item set to find the transitive association rules. Since the join item of a transitive association which have low support value tends to be removed from the next candidate itemset generation, we used two kinds of support values, ‘minimum support’ and ‘minimum relative support,’ in finding the transitive relations. This way the items carrying a strategic importance are given a second chance so that they might be used in discovering the transitive association rules. Our experiments with a real world database showed that the TSAA algorithm was useful in producing transitive itemsets which could not have been obtained otherwise. Observation from other transaction database indicated that there is a ‘minimum relative support level’ at which the number of transitive association itemset decreases suddenly.