Research Article

Accelerating Relevance-Vector-Machine-Based Classification of Hyperspectral Image with Parallel Computing

Table 7

Characteristics of the parallel implementations.

pRVM-MatOppRVM-MultiClasspRVM-Ensemble

Scale of the training setSmallSmallLarge
Explicitly parallelizedNoYesYes
Easy to useYesNoNo
Load balanceYesNoYes
ScalabilityPoorPoorGood
Parallel efficiencyLowLowHigh