Report on the International Workshop on Pattern Representation and Management (PaRMa'04)

摘要

The lack of database techniques supporting the management of patterns has been the main motivation behind the PaRMa'04 workshop. We believe that the papers presented along with the panel discussions have contributed towards this research direction. We now summarize the main outcomes of the workshop: 1. Although it appears too hard to obtain a generic model and language for all kinds of patterns, it is worthwhile to explore this possibility even for a broad subset of cases (e.g., data mining results). Approaches towards the logical and conceptual modeling, as well as the querying of patterns present a valid research challenge. 2. The discovery of pattern similarity and pattern visualization appear to be a couple of approaches that are highly interesting. Both approaches aim at extracting useful knowledge at the pattern space through different means and provide interesting research topics, as well. 3. There are several other research opportunities, with the building of systems and the exploration of their internal operations and data representation mechanisms being the most prominent one. Returning to our starting point, the provision of database-like support to the management of patterns is an open, novel and wide area of research.