ПРЕДСТАВЛЕНИЕ НЕОПРЕДЕЛЕННОСТИ НА УРОВНЕ ТЕНЗОРНЫХ МОДЕЛЕЙ ДАННЫХ

The theory of fuzzy sets, which is an effective tool for solving problems in conditions of uncertainty, in some cases may have certain limitations, due to the complexity of the technological process (or control object), high dimensionality and super-huge volume of the original data set, aggravated by data, contradictory requirements for the membership function, etc. A method for constructing an ordered pairs subset is proposed. This method can be used instead of using the membership function in a case when it is impossible to form it when solving a problem under conditions of uncertainty or as the possibility of obtaining a preliminary solution, which makes it possible to evaluate the rationality of an expertly formed fuzzy set (membership function). The construction of the ordered pairs subset is based on a fuzzy-set interpretation of the singular decomposition of the original data set transformed into a 2D tensor. Examples illustrating the effectiveness of the proposed method are given.