A method for forecasting a nonequidistant (irregular) time series with an irregular sampling interval is presented. Data presented as irregular time series are often encountered in various fields, such as healthcare, biomechanics, economics, climatology, and others. Forecasting irregular time series is in demand in these fields for early warning and proactive decision-making, but there is no universal method for taking into account the unevenness of sampling in the forecast, which determines the relevance of research in this area. The purpose of the study was to develop a method for forecasting a nonequidistant series based on deep neural networks, which allows for good forecast accuracy with a relatively lightweight network architecture. The novelty of the research results lies in the developed method for forecasting nonequidistant time series, the architecture of the deep neural network, and the algorithm that implements the proposed forecast method. The method uses a closed loop, in which the forecast results at the current step are used at the following steps. The original feature of the proposed forecasting method is the use of a multilayer perceptron to forecast the duration of the next irregular sampling interval. This interval is calculated taking into account the correlation time calculated based on the autocovariance function of the durations of irregular sampling intervals. A distinctive feature of the proposed architecture is the presence of a separate input channel of neural network data for analyzing the values of sampling intervals, which allows forecasting the next value of the series taking into account the duration of the forecasted sampling interval. The method is developed for a one-dimensional series, but it can be extended to multidimensional series if the synchronicity of the sampling of the components of the series is observed. The computational experiments showed that with low requirements for computing resources, the accuracy of the forecast based on the proposed method is comparable to modern forecast models within the correlation interval.

The subject of the research is models and algorithms of calendar planning of projects of the target integrated program of ecological rehabilitation of the region, the peculiarity of which is the absence of technological links between program projects. The purpose of the work is to form an economic and mathematical model for determining the rational deadlines for the implementation of projects of the target program for environmental rehabilitation of the region in the absence of technological dependencies between projects; comparing and improving project scheduling algorithms. The result of the work is the formation of an economic and mathematical model that includes two criteria – minimizing the duration of the target comprehensive program and maximizing the progressiveness of achieving its goals, as well as a system of restrictions on the annual amount of investment and the relationship between the desired dates for the start and end of program projects. Two well-known variants of the algorithm for sequentially assigning projects to the calendar plan and a modification of the algorithm are considered, which makes it possible to form the optimal set of projects for each year of the program implementation. The proposed modification of the algorithm using the solution of the problem of finding the optimal set of criteria makes it possible to satisfy both criteria of the economic and mathematical model. To test the operability and effectiveness of the analyzed alternative algorithms, a software package in VBA-Excel was developed. Numerical calculations of the application of the developed algorithm are presented, showing the advantage of the developed algorithm. Conclusions are drawn about the expediency of using this algorithm and the possibility of its adjustment to consider the technological relationships between software projects, which makes it possible to significantly expand the scope of its application.

Probabilistic models for forecasting and assessing the reliability of navigation parameters in intelligent transportation systems are proposed. The relevance of the study is driven by the need to enhance the reliability of robotic transportation systems operating in dynamically changing urban environments. In such environments, sensor failures, signal distortions, and a high degree of data uncertainty are possible. The proposed approach is based on the application of probabilistic analysis methods and statistical control to detect anomalies in navigation parameters such as coordinates, speed, and orientation. The concept of navigation data reliability is introduced as a quantitative measure characterizing the degree of correspondence between the measured parameters and the actual state of the system. Key validity criteria are defined: confidence probability, significance level and confidence coefficients. To improve the reliability of parameter assessment, a combination of statistical analysis methods and filtering algorithms is proposed. Forecasting involves preliminary data processing aimed at smoothing noise and verifying data consistency. Outlier detection is performed using statistical methods, including confidence intervals and variance minimization. An forecasting model based on the Kalman filter and dynamic updating of probabilistic estimates has been developed. The integration of various methods into a unified system minimizes the impact of random and systematic errors, ensuring more accurate assessment of navigation parameters. The proposed approach is applicable to the development of navigation systems for autonomous robots and unmanned vehicles, enabling them to adapt to external conditions without the need for precise a priori data.

In this paper we propose a method for analyzing incomplete and inaccurate data in order to identify factors for predicting the volume of mudflows. The analysis is based on the mudflow activity inventory data for the south of Russia, which is poorly formalized, has missing values in the mudflow types field, and requires significant additional processing. Due to the lack of information on the mudflow type in the cadastral records, the primary objective of the study is to develop and apply a methodology for classifying mudflow types to fill in the missing data. For this purpose, a comparative study of machine learning methods was performed, including neural networks, support vector machines, and logistic regression. The experimental results indicate that the neural network-based model has the highest prediction accuracy among the methods considered. However, the support vector machine method demonstrated a higher sensitivity rate for classes represented by a small number in the test sample. In this regard, it was concluded that an integrated approach is appropriate, combining the strengths of both methods, which can help improve the overall classification accuracy in this subject area. Forecasting the volume of material removal and data clustering showed the presence of nonlinear dependencies, incompleteness and poor structuring of data even after filling in missing values of the mudflow type, which required a transition from numerical data to categorical data. This transition increased the model’s resistance to outliers and noise, allowing for a highly accurate forecast of a one-time removal. Since the forecast does not reveal the factors influencing its result, an analysis was conducted to identify these factors and present the found patterns in the form of logical rules. The formation of logical rules was carried out using two methods: the associative analysis method and the construction of a logical classifier. As a result of applying associative analysis, rules were found that reflect some patterns in the data, which, as it turned out, need significant correction. The use of the developed logical methods made it possible to clarify and correct the patterns identified using associative rules, which, in turn, ensured the determination of a set of factors influencing the volume of the mudflow.

All steganographic methods are focused on a specific container file format. Text documents with markup are the most difficult object for steganography methods. The article suggests a model for embedding structured text documents in control tags. The model uses the document tree structure and embeds into free leaf nodes. This approach adds hidden data that does not affect the display of the document. Two steganographic methods are implemented based on this model. The first method embeds hidden data into html document tags. The embedding method adds underplayed tags and style classes to free leaf nodes. The hidden data extraction method uses the embedding identifier. This role is played by the name of the new class. The name generation algorithm is based on the embedding key and hash function. The format of the identifiers matches the format of the source document names. This naming method allows the hidden message blocks to be randomly allocated to free leaf nodes. The second method embeds steganographic inserts into xml documents. Hidden data is added to the free leaf node attributes. The method requires two new attributes to execute. The optional structure describes both attributes. The format of this structure is indistinguishable from the structures present in the document. The embedding identifier is also based on the embedding key and the embedded block number. The data view uses an encryption algorithm with an additional key. Both methods use embedded data masking to counteract source code steganalysis. Steganalysis of such methods has exponential algorithmic complexity, so both methods are only applicable to large files.

The effectiveness of management activities is largely determined by the degree of reasonableness and efficiency of selecting a decision-maker and delegating the necessary powers to him. To solve this problem, it is advisable to use mathematical methods and information technologies that ensure the specified selection in an automated mode. This approach involves the creation of a DSS using unpopulated decision-making technologies. This allows you to identify employees or groups of employees who are most suitable for the implementation of managerial functions in terms of certain quality indicators. The use of unpopulated decision-making technologies in the interests of economic entities makes it possible to reduce the time and cost of their adoption by automating the transit management level. The purpose of the study considered in this article is to solve the problem of organizing the effective selection of a decision maker (DPR), or a group of people, as well as transferring the task to be performed automatically. To achieve this goal, the mathematical modeling of the LPR selection algorithm (LPR group) was carried out by performing operations on sets containing significant indicators and solving a mathematical programming problem. The proposed mathematical apparatus makes it possible to create a subsystem of the DSS responsible for the selection of a DSS (group of DSS) capable of making a high-quality management decision under given constraints, and transferring tasks to it for execution. The specified restrictions contain the type of task to be solved, the possibility of its implementation with the help of unpopulated decision-making technologies, competence limitations of LPR (groups of LPR), time constraints. The DSS subsystem, created on the basis of the proposed mathematical apparatus containing the task of optimizing the selection of LPR (LPR group) according to the cost criterion, makes it possible to minimize the cost of decision-making.

The current period of technological progress is characterized by a significant penetration of digital technologies into all spheres of life and society. At the moment, a fairly extensive material has been accumulated with the results of research aimed at recognizing the effects, mostly implicit or weakly exposed, which are inherent in digital technologies, as well as to disclose the mechanisms producing them. However, work in which the task of creating a holistic representation would be solved, reflecting the most important effects in their entire entirety, have not yet been published. The article attempted to analyze all the main features of the application of such technologies in an academic environment, closely related to information processes and where the introduction of digital technologies is especially active. In order to summarize theoretically and practically proven strengths of digital technologies, to identify the most successful applications built on their base, to identify negative manifestations and effective steps taken to neutralize threats, a fairly representative set of research results on identifying individual factors and the study of the effects due to them, specific for digital technologies for a certain applied orientation. Based on the basic concepts, taking into account the experience of the implementation of organizational and administrative measures, focused on achieving the most positive effect of the introduction of digital technologies and elimination of sources of undesirable consequences, a set of key objects of comprehensive analysis is proposed, which should precede decision-making to integrate technologies into the practical activities of the academic structures and serve as the basis for the formation of general policy and strategic plan.

The article considers a fundamentally new generation of soft exoskeletons based on magnetic rheological modules. A model of an exoskeleton link with adjustable rigidity is proposed, for which a system of differential equations of motion is compiled. A model of an exoskeleton for human lower extremities, composed of modules with magnetic rheological fluid, has been developed. The problems of kinematics and dynamics that arise when controlling the movement of an exoskeleton are presented. A block diagram of an algorithm for controlling the movement of an exoskeleton has been developed, solving control problems taking into account neural network methods. A qualitative study by analytical methods of the proposed mathematical model of an exoskeleton with links with adjustable stiffness is carried out. The use of regulation of the rigidity of the links of the exoskeleton modules in the necessary way will make it possible to create devices that are comfortable for the user, both from the physiological and psychological points of view. Soft exoskeletons will be in demand in many industries and medicine. The use of electromagnetic actuators, which, as a rule, have smaller dimensions and energy consumption, an almost instantaneous response speed in comparison with pneumatic drives, hydraulic drives and electric drives, is promising and technically, energetically, environmentally, economically more acceptable. Under the action of electric or magnetic fields, the stiffness of electroelastic and magneto-rheological materials can change by more than an order of magnitude. This makes it possible to create exoskeleton links with adjustable stiffness. Regulation of the rigidity of the elements of the exoskeleton modules in real time is a complex control task. Therefore, exoskeleton motion control requires the development of adaptive type algorithms based on a neuro-fuzzy approach, using information received from feedback. Neural network algorithms for self-learning in the process of functioning are a logically expedient part of a soft bionic exoskeleton. Also, as a necessary component of the exoskeleton, along with the actuator part, a sensor system is needed with sensitive sensors for the user’s neuromuscular impulses and various sensors for position, speed, acceleration, force, moment of force of the exoskeleton. A comprehensive study of scientific and technical problems of the functioning of the model of the magnetic-rheological medium, exoskeleton modules and their compilation into a single multi-link structure, its movement using the technology of modern mathematical modeling was carried out. The construction of neuro-fuzzy control systems for soft exoskeletons is described as part of the development of new mathematical methods for modeling complex technical objects using modern mechatronics and cyborgics.

During the analysis of the most common architectures for building neural networks for image analysis – a direct convolutional layer, a recurrent neural network, a convolutional neural network – the authors determined that the last architecture option for the task is the most suitable. The solutions available on the market, although they allow us to recognize emotions, do not evaluate them, and this is an important result for a wide variety of tasks, because such an assessment allows us to more accurately predict the possible future behavior of both individuals and groups of people. The purpose of this article is to design an application software product (application) and develop a prototype that could perform the functions of analyzing emotions and evaluating them. The objectives of this study include the choice of the architecture of the designed application; the development of an algorithm for the operation of the application; the design of the user interface; a description of the neural network learning process and its results, the model of which was defined in the previous article; demonstration of the prototype (control example). The scientific novelty of the projected application lies in the formation of an assessment of the psycho-emotional state of those people whose images of their faces were evaluated. Such an assessment with recommendations can be widely demanded in various branches of human activity, since it is important to be able to express emotions not only to ordinary people, but also to those whose job is to transfer their emotional state to others: business coaches, journalists, actors, animators, dancers, etc. The presence of emotional intelligence is very important at the moment time.