The Gradient Method of Constructing an Image of Mapping for the Classification Problem of the Object Operation Modes

The paper describes the solution of the problem of discrimination (classification) of object operation modes by construction of areas corresponding to each of the considered modes types, with subsequent verifica-tion of falling points, defined by the observed states of the object, in one of the areas. As a source of infor-mation about an object, its simulation model and acceptable ranges of object parameters are used. The map-ping under study combines the simulation model function and the gauging function. All this makes it possible to turn the task of training the classifier to the task of finding the image of the mapping. The proposed approach to the solution of the problem is based on the maximization of the informativity of points, approximating the boundary of the area. The area of the required domain and the absolute deviation of the lengths of the edges forming the boundary of the domain from their average value serve as the informativity criterion. The proposed approach is applied to solve the problem of differentiating the operation modes of an object by constructing an image of the mapping and finding its boundary line. The results of the algorithm are analyzed, its advantages and disadvantages are presented. Obtained images of mappings can be used as the basis of a classifier of ob-ject operation modes. The optimization procedure in this case will be a training procedure for the classifier.