Mining Interesting Purchase Patterns: A Method of Granular Computing

Nowadays, information technologies are widely used in business. Enterprises hope to make good use of the advanced technologies to analyze customer purchase behavior for better marketing. So it has become a hot issue to find interesting customer purchase patterns in a large amount of information. This paper proposes a method of granular computing to mine interesting purchase patterns. The granule represents a set of tuples that have the same attribute value in the database. For all tuples involved in interesting purchase patterns, we can represent them by using existing granules or creating new granules by logical operations (AND or OR) among existing granules. And then we use these granules to generate interesting patterns. This method not only can improve performance efficiently without scanning database repeatedly, but it is easier to understand for users and improves process flexibility. Especially for some complicated user demands, some high-level concepts which users take interest in don't exist in the database, but we can generate new granules by OR operations to represent them. So it's very convenient to mine interesting purchase patterns using granular computing.