Multi-Cluster Text Mining on the Grid using the D-Grid UNICORE environment

Text mining is inherently more computation-intensive than information retrieval on pre-structured data, and it requires transfer and filtering of huge amounts of data. Grid environments provide a suitable infrastructure for ac- complishing these tasks. We present the mapping and implementation of a standard text mining workflow for the analysis of biomedical text data from PubMed to a D-Grid UNICORE environment with multiple PC-clusters. We, furthermore, discuss the gain in applicability, the open is sues of our solution and possible future enhancements.