Mining of both interesting positive and negative association rules based on MLMS model

Abstract

A variety of data in data mining can be an open gate to extensive studies of various aspects in business. This research was aimed to analyze the data mining association rules to find ways for improving process efficiency. This would be of great benefits to people involved in commercial transactions in terms of competitive advantages of information data and the inventory development. In this research, we studied using pruning strategy of both positive and negative association rules with only single minimum support. In addition, frequent itemsets and infrequent itemsets in the database were focused. Further study was on the work which multiple levels of minimum supports in relation to the length of itemsets were employed. From the study, the application of a prune strategy, together with minimum supports to the multiple levels had resulted in a new technique which helped reduce the process of the association rules of interest, both positive and negative on the MLMS model.