Collaborative management of a repository of KDD processes

Abstract

Knowledge Discovery in Databases (KDD) is a complex and computationally intensive process that requires a repeated interaction between tools and users, often in a distributed environment. Given the complexity of the process, both naive and expert users need some support to effectively perform knowledge discovery. In this paper, we present a user-and knowledge-centric approach to support the design of KDD projects. Semantic technologies are exploited to support sharing and (re)use of KDD computational resources and processes in a distributed collaborative environment. In particular, functionalities for tool publishing, service and process discovery, and versioning of processes greatly enhance process management, and provide a learn-by-example and trial-and-test environment for collaborative KDD design. The systematic use of semantic information, a loosely coupled and layered service-oriented architecture and a cooperative and flexible approach results in a platform natively conceived for an open, distributed and collaborative environment.