Frequent and Significant Patterns Mining Using Approximate Patterns for Protein Structure Analysis

Abstract

With the advent of technology and tools, large volumes of data have generated in varying complex forms. Graphs and graph based data mining has emerged as an appropriate solution to represent and to mine useful knowledge from such complex structured data. In order to extract useful information from these datasets, many algorithms are being developed for various graph based data mining tasks such as frequent pattern mining, classification, clustering and indexing in graph data. Graphs are especially appropriate to model proteins structures and to determine the structural and functional characteristics of different proteins which is evolving as a key area in genetic information processing. In this paper, we present FSPM, a novel framework which retrieves both frequent and significant patterns directly at a time using approximate patterns. To the best of our knowledge this is the first work that mines frequent and significant patterns at a stretch. Our preliminary experiments demonstrate the efficiency of the proposed framework.