Heuristic approach to model of corporate knowledge construction in information and analytical systems

Abstract

Concerning intelligent information and analytical systems one of the most prospective areas is the construction of knowledge bases used ontological systematization as a tool for classification of corporate knowledge. The authors interpret a competence model, which can select significant features of classifiable objects, in terms of domain ontology. To classify corporate objects it is suggested a heuristic method of knowledge clusterization in multidimensional feature space in which a genetic algorithm is used to obtain effective solutions for classification procedure according to well-known criteria. The genetic algorithm is an iterative probabilistic search algorithm whose main feature is simultaneous using of a set of population from the space of potential solutions. A certain advantage of the method is guaranteed lack of intersections for all clusters and necessary to define the number of clusters. Experimental results were carried out on the basis of test tasks and confirmed a theoretical relevance and promising of the suggested method.