An OLAM-Based Framework for Complex Knowledge Pattern Discovery in Distributed-and-Heterogeneous-Data-Sources and Cooperative Information Systems

Abstract

The problem of supporting advanced decision-support processes arise in many fields of real-life applications ranging from scenarios populated by distributed and heterogeneous data sources, such as conventional distributed data warehousing environments, to cooperative information systems. Here, data repositories expose very different formats, and knowledge representation schemes are very heterogeneous accordingly. As a consequence, a relevant research challenge is how to efficiently integrate, process and mine such distributed knowledge in order to make available it to end-users/applications in an integrated and summarized manner. Starting from these considerations, in this paper we propose an OLAM-based framework for complex knowledge pattern discovery, along with a formal model underlying this framework, called ${\mathcal M}ulti-Resolution ~ {\mathcal E}nsemble-based ~ Model for Advanced ~ {\mathcal K}nowledge~ {\mathcal D}iscovery ~in ~Large ~{\mathcal D}atabases~ and~ Data~ Warehouses$ $\mathcal{MRE-KDD}$ ⁺), and a reference architecture for such a framework. Another contribute of our work is represented by the proposal of KBMiner, a visual tool that supports the editing of even-complex KDD processes according to the guidelines drawn by $\mathcal{MRE-KDD}$ ⁺.