On Service Discovery for Online Data Mining Trails

Complicated data mining tasks require not only data and computation intensive facilities but also non-trivial trails of various algorithms and their parameters, which can be best exploited through a collaborative way in virtual communities. This paper proposes a service discovery system to tackle these requirements in a grid enabled online data mining environment. The system adopts the state-of-the-art open standards to represent data mining algorithms and models. Then different service discovery mechanisms are proposed. Quality and performance of proposed strategies demonstrate its usefulness for online data mining trails.