First Order Dynamic Instance Selection
Peter Géczy, Shiro Usui, J. Chmúrny · 2000
Training of adaptable systems such as neural networks indispensably depends on the training exemplar set. The most promising training algorithms utilize dynamic instance selection. Dynamic instance selection technique is capable of selecting instances dynamically at each iteration of adaptation procedure. Adaptable system is thus at each iteration presented with appropriately selected set of learning instances that can vary in size and content. Variability of the selected exemplar set contributes to the speed of learning and lowers its computational cost. Benefit of dynamic instance selection can also be found in improved properties of trained adaptable systems. These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.