Similarity Association Pattern Mining in Transaction Databases

Abstract

Association pattern mining is a method of finding interesting relationships or patterns between item sets present in each of the transactions of the transactional databases. Current researchers in this area are focusing on the data mining task of finding frequent patterns among the item sets based on the interestingness measures like the support and confidence which is called as Frequent pattern mining. Till date, in existing frequent pattern mining algorithms, an itemset is said to be frequent if the support of the itemset satisfies the minimum support input. In this paper, the objective of our algorithm is to find interesting patterns among the item sets based on a Gaussian similarity for an input reference threshold which is first of its kind in the research literature. This study is limited to outlining naïve approach of mining frequent itemsets which requires validating every itemset to verify if the itemset is frequent or not.