Some Selected Soft Computing Tools and Techniques
Bożena Kostek · Studies in fuzziness and soft computing · 1999
There are several definitions concerning soft computing as a domain of science. The most widely known and most often applied soft computing (or computational intelligence) methods are neural networks, multivalued logic, fuzzy sets and fuzzy logic [219], Dempster-Shafer theory [215], rough sets [161], probabilistic reasoning, evolutionary computation, etc. Particular attention was paid in this work to neural networks, fuzzy logic and rough sets. Neural networks may be treated as tools for modeling dependencies between variables. On the other hand, both fuzzy and rough sets are formal methods for dealing with uncertainty. These techniques are reviewed further in this chapter, because they are used to provide a kernel to decision algorithms as applied to classification tasks. A particular justification for the application of decision systems in this area is provided by the fact that the management of uncertainty in acoustics should be based on the knowledge of experts — the best criterion for assessing the acoustical quality of music. 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.