卷 19, 编号 6 (2020)

The recommendations on the application of methods of multidimensional estimation (MDE) of objects, proposed in the paper Velasquez M., Hester P.T. «An Analysis of Multi-Criteria Decision Making Methods», are analyzed. The weak substantiation of these recommendations, resulting from the superficial systematization of MDE methods, is noted. The recommendations are focused not on the classes of MDE methods, but on various areas of activity. However, in each area of activity there is a wide range of tasks for evaluating objects of various nature. In this regard, the urgency of a more thorough systematization of MDE methods is recognized. Taking into account the diversity of MDE methods, it was decided to limit ourselves to the systematization of methods that use evaluation functions (EF), and on this basis to offer general recommendations for their application. The review of MDE methods from a unified position required clarification of the terminology used in them. On the basis of the formal model of the criterion, the relationship between the concepts of "preference", "criterion" and "indicator" is established. To highlight the methods that use evaluation functions, the concept of the target value of the indicator is introduced. Regarding its location on the indicator scale, the concepts of ideal and real goals are introduced. The criteria corresponding to these goals are divided into target and restrictive ones. Using the proposed terminology, a review of the most well-known MDE methods was carried out. Of these, a group of methods using evaluation functions is distinguished. Variants of evaluation functions created on the basis of the criterion and postulates of the theory of value and utility are considered. On the basis of the similarity of the domains of definition and the meanings of EFs, the relationship between them is established. Regarding the target value of the indicator, they are divided into the functions of achieving the goal and functions of deviation from the goal. The mutual complementarity of these functions is shown. A group of functions of deviation from the goal is highlighted, which allows us to order objects separately according to penalties and rewards in relation to achieving a real goal. The concept of norm is introduced for the correspondence relation. On the example of medical analyzes, the practical application of deviation functions from the norm is shown using both the minimax and the weighted average generalizing function to establish a rating on a set of objects. The similarities and differences of the EFs revealed in the course of the study form the basis for the classification of the MDE methods that use them. The difference in EFs in terms of the complexity of creation is reflected in the proposed methodology for their application.

A problem of searching a minimum-size feature set to use in distribution of multidimensional objects in classes, for instance with the help of classifying trees, is considered. It has an important value in developing high speed and accuracy classifying systems. A short comparative review of existing approaches is given. Formally, the problem is formulated as finding a minimum-size (minimum weighted sum) covering set of discriminating 0,1-matrix, which is used to represent capabilities of the features to distinguish between each pair of objects belonging to different classes. There is given a way to build a discriminating 0,1-matrix. On the basis of the common solving principle, called the group resolution principle, the following problems are formulated and solved: finding an exact minimum-size feature set; finding a feature set with minimum total weight among all the minimum-size feature sets (the feature weights may be defined by the known methods, e.g. the RELIEF method and its modifications); finding an optimal feature set with respect to fuzzy data and discriminating matrix elements belonging to diapason [0,1]; finding statistically optimal solution especially in the case of big data. Statistically optimal algorithm makes it possible to restrict computational time by a polynomial of the problem sizes and density of units in discriminating matrix and provides a probability of finding an exact solution close to 1. Thus, the paper suggests a common approach to finding a minimum-size feature set with peculiarities in problem formulation, which differs it from the known approaches. The paper contains a lot of illustrations for clarification aims. Some theoretical statements given in the paper are based on the previously published works. In the concluding part, the results of the experiments are presented, as well as the information on dimensionality reduction for the coverage problem for big datasets. Some promising directions of the outlined approach are noted, including working with incomplete and categorical data, integrating the control model into the data classification system.

This paper presents a comparison between discrete Hidden Markov Models and Convolutional Neural Networks for the image classification task. By fragmenting an image into sections, it is feasible to obtain vectors that represent visual features locally, but if a spatial sequence is established in a fixed way, it is possible to represent an image as a sequence of vectors. Using clustering techniques, we obtain an alphabet from said vectors and then symbol sequences are constructed to obtain a statistical model that represents a class of images. Hidden Markov Models, combined with quantization methods, can treat noise and distortions in observations for computer vision problems such as the classification of images with lighting and perspective changes.We have tested architectures based on three, six and nine hidden states favoring the detection speed and low memory usage. Also, two types of ensemble models were tested. We evaluated the precision of the proposed methods using a public domain data set, obtaining competitive results with respect to fine-tuned Convolutional Neural Networks, but using significantly less computing resources. This is of interest in the development of mobile robots with computers with limited battery life, but requiring the ability to detect and add new objects to their classification systems.

The high spectral efficiency of signals with continuous phase modulation (CPM) has determined their popularity and active use in various radio engineering projects. The uniqueness of the properties of CPM signals is associated with the preservation of the continuity of their phase when changing information messages for the duration of a symbol. At the same time, until recently, of the entire wide class of signals with continuous phase modulation, the most widespread were various variations, the so-called Minimum Shift Keying (MSK) signals. However, these are far from the only representatives of the class of CPM signals with the property of high spectral compactness. This article examines no less interesting signals of this class, formed by means of Dual Phase Modulation (DPM). In particular, analytical expressions of their synthesis are presented, their belonging to the class of CPM signals is substantiated. In addition, the article investigates the temporal properties of the phase function recommended by ITU-R SM.328-11 for the synthesis of signals with continuous phase modulation, presents the time and frequency fragments of MSK signals in comparison with signals with Binary Phase Shift Keying (BPSK). The stages of the analytical derivation of the model of noise immunity of PCM signals in terms of the probability of a bit error based on an empirical approach are presented. The generality of the obtained model with the known expression for MSK signals is shown by studying the difference function of the approximation error (error of the order of 10-3), which made it possible to obtain a more compact representation of the developed model in relation to DPM signals. It has been proven that DPM signals have higher noise immunity properties in relation to MSK signals (about 0.5 dB at an error level of 10-5), using the results of studying the difference functions determined by the difference between the signal symbols corresponding to the information values "1" and "0". The directions of further research are determined.

Currently, centrally reserved access to the medium in the digital radio communication networks of the IEEE 802.11 family standards is an alternative to random multiple access to the environment such as CSMA/CA and is mainly used in the transmission voice and video messages in real time. Centrally reserved access to the environment determines the scope of interest in it from attackers. However, the assessment of effectiveness of centrally reserved access to the environment under the conditions of potentially possible destructive impacts was not carried out and therefore it is impossible to assess the contribution of such impacts to the decrease in the effectiveness of such access. Also, the stage establishing of centrally reserved access to the environment was not previously taken into account. Analytical model development of centrally reserved access to the environment under the conditions of destructive influences in digital radio communication networks of the IEEE 802.11 family standards. A mathematical model of centrally reserved access to the environment has been developed, taking into account not only the stage of its functioning, but also the stage of formation under the conditions of destructive influences by the attacker. Moreover, in the model the stage of establishing centrally reserved access to the medium displays a sequential relationship of such access, synchronization elements in digital radio communication networks and random multiple access to the medium of the CSMA/CA type. It was established that collisions in the data transmission channel caused by destructive influences can eliminate centrally reserved access to the medium even at the stage of its establishment. The model is applicable in the design of digital radio communication networks of the IEEE 802.11 family of standards, the optimization of such networks of the operation, and the detection of potential destructive effects by an attacker.