Strategies for Parallelizing Data Mining

@inproceedings{Skillicorn2007StrategiesFP,
  title={Strategies for Parallelizing Data Mining},
  author={David B. Skillicorn},
  year={2007}
}
We classify parallelization strategies for data mining algorithms, concentrating on those techniques in which training data is partitioned, and extracted properties shared between processors at the end of each phase. This approach has been extensively investigated for association rules and decision trees. We sketch some similar work for supervised and unsupervised neural networks, and for the kDNF technique. Possibilities for more sophisticated, hybrid, algorithms are also indicated. Parallel… CONTINUE READING