Introduction to Meta‐Algorithmics
Steven J. Simske · 2013
Collectively, meta-algorithmics are called intelligence generators, and the author alludes to them as such, except where he explicitly calls out algorithms, systems, or large engines such as optical character recognition (OCR) or automatic speech recognition (ASR) knowledge engines. In this chapter the three different orders of meta-algorithmic patterns are introduced. A large—but not exhaustive—set of patterns are provided for each of these three levels. Several of the powerful tools for meta-algorithmics—confusion matrices, weighting, output space transformation, sensitivity analysis, and feedback—are explained. Having provided the patterns—which like all patterns are “templates” for domain-specific implementation—we next turn to application of these patterns to a wide, but again not exhaustive, set of relevant intelligent systems. Controlled Vocabulary Terms ambient intelligence; optical character recognition; speech recognition