A methodical approach for mining negative rules

Abstract

Discovering hidden rules and facts from the database is one of the most important aspects in knowledge discovery. Our main focus in this paper is to find the negative association rules based on a classical approach. Negative association rules are useful in Market Basket Analysis to identify the conflicting products (complementary products).We have used apriori algorithm as a base to present a new algorithm to find negative association rules. Our algorithm generate non frequent sets categorized
as weak non frequent and strong non frequent based upon a minimal support. The algorithm is based on the approach that superset of every strong non frequent subset is always strong non frequent.