Integrating Pattern Mining In Relational Databases (original) (raw)

This paper presents a novel approach to pattern mining in relational databases by integrating the concept of virtual mining views. Instead of requiring users to interact with complex mining algorithms directly, the proposed system allows users to query patterns as if they were part of a traditional database management system. Through triggering mining algorithms transparently upon user queries, the system efficiently manages the large volume of potential mining results without the need to store them physically. A specific focus is placed on the constraints extraction process, enabling users to filter patterns effectively.