VizSFP: A visualizer for semantically similar frequent patterns in Dynamic Datasets

Frequent pattern mining is an important area in data mining. Transactional databases are insufficient to analyze current shopping trends and as such we should consider Dynamic Datasets that updates transactions in an adhoc basis. Algorithms such as Apriori require more time to generate huge set of rules. Discovering interesting rules from the generated rules is difficult. Works that are reported until now in reducing number of rules are either time consuming or does not consider the interestingness of the user and does not focus on analysis of rules. This paper presents a case study on grocery data that uses SSFPOA semantic measure to reduce number of generated patterns, clusters the similar patterns and visualizes these clusters for easy analysis. Six graphs namely NCGraph, NSGraph, LCGraph, LSGraph, NEGraph and HGraph are proposed in VizSFP for visualizing frequent patterns. Clusters that are formed by SSFPOA are validated using clustering validating indices.