The formation of a characteristic space in classification problems can be divided into two stages: the choice of the initial description of objects and the formation of an informative description of objects on the basis of a reduction in the dimension of the space of the original description.

The definition of an informative set of features is one of the important tasks in pattern recognition. Typically, the determination of informative features is carried out using two types of methods. The first type of methods is “direct methods”, they are directly aimed at identifying informative sets of attributes. And the second type is called "inverse methods", these methods serve to build informative sets of signs by eliminating uninformative signs from the attribute space. This article is devoted specifically to the development of the second type of method; it proposes an accelerated method and an algorithm for determining non-informative features based on the selected non-informative criterion.

In solving most of the practical problems associated with classification, a heuristic approach is widely used. Currently, many heuristic methods and algorithms based on various heuristic criteria have been developed. The most popular criteria are heuristic informational criteria. These criteria are related to the assessment of the separability of these classes and are based on the fundamental hypothesis of recognition compactness, i.e. with an increase in the relationship between classes, their separability improves. If traits maximize relationships, they are called "good." Such heuristic criteria are widely used and give good results in solving practical problems, but they are poorly studied in theoretical terms. At present, the method for choosing non-informative features, taking into account the relationship of features based on heuristic criteria, has not yet been developed. The article considers the problem of determining informative features by eliminating uninformative ones.

The solution of many applied problems is reduced to solving optimization problems with the objective function of a linear-fractional functional of “-1” order. For example, economic tasks, tasks of placing objects, tasks about assignments, layout, production planning, an informative description of objects, medical tasks, etc. The article developed a new method for solving the unconditional optimization problem with the objective function of the discrete-fractional-linear functional of the “-1” order. As well as certain optimality conditions for the chosen solution, a theorem is formed and proved, which is the basis of the proposed method.

At present the most popular criteria are of informative heuristic criteria associated with the estimation of separability given classes and based on the fundamental pattern recognition compactness hypothesis: with increasing distance between the classes improved their separability. “Good” are those features that maximize the relationship. Such heuristic criteria, although are widely used in solving practical problems of classification, but in theory are scarcely explored. At present, the method of selecting informative features, taking into account the relationships of features based on heuristic criteria, has not been developed. The report considers this task.