Association Mining in Borderline Itemsets
We present a modification to the association mining algorithm Apriori. The goal of our modification is to determine what influence, if any, frequent itemsets have on borderline itemsets. The purpose of identifying this influence is to look for possible ways to increase the frequency of the borderline sets. An example of a practical application would be to increase sales of a poorly selling product instead of simply dropping it from the inventory.
The AprioriLF algorithm is used to identify the borderline cases when generating item- sets. Tests were run to determine the correctness of the base Apriori algorithm. Then more tests were run on the extended algorithm to examine the grouping of each frequent itemsets and the borderline itemsets.