Pattern mining for historical data analysis by using MOEA

Abstract

In data mining, graph mining is a promising new approach and some algorithms are proposed. However, their application is limited in the field of business. This is because of the wide diversity of business data. In this paper, we propose a method which extracts new valuable patterns by using graph mining approach and MOEA. In our method, historical purchasing data for each customer is transformed into tree structured data and gene is constructed from the structured data at first. Then the patterns are extracted by using existing MOEA from these genes. We apply our proposed method to a practical business data. From computational experiments, we show that our method has good performance and is able to extract valuable patterns from the view of business. © 2009 Springer Berlin Heidelberg.