✍ Scribed by Vladimir Jotsov; Vassil Sgurev
No coin nor oath required. For personal study only.
Contemporary knowledge discovery systems are mainly quantitative. This part of the research could be successfully combined with the considered qualitative research in acquisition, elicitation, and discovery of logic-based rules and patterns. The paper introduces a synthetic metamethod (SMM), which is aimed to support different types of creative processes. SMM is applicable in quite different fields of computational discovery, knowledge discovery and data mining, information security, number theory, and computer science but the paper focuses on its applications in intelligent systems. Previous research shows that the main part of the considered set of methods is domain independent and is applicable to a wide range of inconsistency tests, conflict resolution, detection or identification, and in fuzzy perceptions approval. Also, it can be used for elimination of weak analogical or other hypotheses or during application of creative processes concerning different types of intelligent systems. The metamethod SMM is based on the analysis of incoming knowledge: beliefs, probabilistic estimations, hypotheses, perceptions, etc., and a successive check with the existing knowledge. The main methods proposed-SMM, INCONSISTENCY, FUNNEL, and CROSSWORD-interact on a competitive principle. The convergence of the obtained results by its character is close to the one from the group of the data-mining methods. The operation is not quite autonomous; it is evolutionary by nature, and admits a kind of "course change" after the human intervention that can be realized at any time. On the other hand, one of the basic goals of the presented results is to prompt the user about possible nontrivial decisions. It may be considered as a new kind of human-machine-brainstorming methods. A sophisticated evolutionary method is applied to knowledge-poor environments or to the missing precise orientation to the final solution.
📜 SIMILAR VOLUMES