Representing Discovered Patterns Using Attributed Hypergraph

Yang Wang, Andrew K. C. Wong

One of the fundamental problems in knowledge discovery in databases and other applications of AI is how to represent knowledge and patterns. Existing representation schemes have various shortcomings. In this paper, we propose a new knowledge representation scheme using attributed hypergraph (AHG), which is simple yet general enough to directly encode different order patterns discovered from large databases. In AHG, both the qualitative and quantitative relations are represented as attributed hyperedges. Such representation is lucid and transparent for visualization. Besides, patterns in AHG are easy to understand. In the discussion, some basic manipulations of AHG for data mining tasks are briefly addressed. The paper ends with examples of pattern representation using AHG.

This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.