Mining Positive and Negative Association Rules in Multi-database Based on Minimum Interestingness

With the increasing development and application of information and communication technologies, multi-database mining is becoming more and more important. Association rules mining is the major topic in multi-database. According to Piatetsky-Shapiropsilas argument, an association rule is interesting only if the rule meets the minimum interestingness condition. In this paper, we extended this condition to mine association rules in multi-database and improved it to check the correlation of association rules. An algorithm PNAR_MDB _on P-S measure is proposed and the experimental results demonstrated the algorithm is effective.

1. Introduction

Database mining has emerged as a major application area for an efficient discovery of the previously unknown and potentially useful patterns in large databases. Traditional association rule mining has been mainly focused on identifying the relationships strongly associated among itemsets that have frequent and high-correlation. It is the form call positive association rules (PARs). As an important supplement, the other three forms call negative association rules (NARs) which can catch mutually exclusive correlations among items, they can provide much useful information and play an important role in decision-making.