Spidal Kernels

This work refines and exemplifies the Ogre facets, which include core machine learning and other kernels, as well as those related to data systems. In order to tie in to previously established Big Data guidelines we set out to catalog existing benchmark sets and label them with corresponding Ogre facets. Kernels identified in SPIDAL are also related to Ogres. Continuing work will serve to identify those facets not covered so far. Reference paper 3 below studies SPIDAL clustering algorithms General Machine Learning (GML) 1 and 2.

  • Research paper here.
  • Research paper here.
  • Research paper here.

Big Data Benchmarking

Below is a list of the members of SPIDAL library and the Ogre facets that they support. There are no SPIDAL library members directly addressing Data Source & Style View (except spatial analytics and GIS) and so that view is omitted.

Ogre Facet Benchmark List

SPIDAL Kernels Big Data Benchmarking Further/Ongoing work:

  1. Package existing Kernels
  2. Examine Intel DaaL/Mahout/MLlib
  3. Document SPIDAL use of Github
  4. Implement missing algorithms
  5. Consider implications of GPUs and Xeon Phi “Knights landing”