Hybrid Distribution for Association Rules Extraction on Grid Computing
Mohammed Rebbah, Miloud Khaldi, Stambouli Mascara · 2015
The extraction of association rules in distributed systems is capable of greatly reducing the time of extraction and exploitation of large data sets, however, these benefits are forced to emerging issues related on the one hand the nature of distributed systems and the sharp increase in volumes of data and its geographic dispersion. The trend towards grid computing that enables applications to handle distributed heterogeneous computing resources as a single virtual machine is justified by these last two points. Globus_HyDAR, our service developed under Globus grid, falls within the framework of grids Data mining; has proposed a hybrid intelligent distribution (horizontal and vertical) data. The results clearly show gains execution time compared to the horizontal and vertical distributions.