| http://www.w3.org/ns/prov#value | - While algorithms that attempt to learn in such environments have recently started to appear in the literature, these algorithms are mostly ad hoc and heuristic in nature with many free parameters to be fine-tuned, make restrictive assumptions that do not generally hold in real world (slow / gradual drift, non-cyclical environments, no new classes, or availability of old data at all times, etc.), a
|
