Modeling Execution Times of Data Mining Problems in Grid Environments



The problem of distributing data mining tasks in Grid environments in order to shorten overall execution times is addressed. The text categorization case study shows that gains heavily depend on data transfer required to distribute jobs. Therefore, simple and
intuitive models of data transfer in Condor and Alchemi Grid environments are presented. In most cases the models reliable estimate the execution times of parallelized tasks.


Text mining; SVM;GRID

Related Project

GRID II - Global GRID for Data Mining with Soft Computing on Large Data Bases


ERK 2005, September 2005

Cited by

No citations found