FORECASTING THE COST OF ELECTRICITY AND THE STATE OF INSULATION OF ELECTRICAL EQUIPMENT
Y.M. Kulinich, S.A. Shuharev
The paper deals with the application of the method of time series analysis to predict the cost of electricity and assess the state of insulation of power circuits of an electric locomotive. The proposed approach makes it possible, on a scientific basis, to plan the amount of funds allocated to pay for electricity, as well as to take timely measures aimed at restoring insulation and excluding the causes of fires that occur on locomotives. Time series analysis was carried out with the help of an application program that allows assessing the trend of changes in the indicators under consideration. A device for monitoring the state of insulation of power circuits of an electric locomotive is also proposed, in which the developed program for forecasting time series is implemented. Installing the device described in the work on the locomotive will allow timely assessing the current and predicted state of insulation, as well as taking timely measures to restore it. The urgency of the problem of diagnosing the state of insulation is due to the aging processes of fixed assets (machine tools and equipment) at industrial enterprises, which requires timely measures to restore the state of insulation of power electrical equipment. The application program is implemented in the MatLab package and is used to predict the cost of electricity. To expand the possibilities of using the application in other applications, the source code of the program was transformed into the code written in the high-level C language. The program obtained in this way is used in the PIC18F452 microcontroller to assess the state of the insulation of the power circuits of an electric locomotive.
Keywords: time series forecasting, singular spectral analysis method, singular decomposition, electricity cost, insulation condition.
CALCULATION OF INDICATORS OF THE INFORMATION CONTENT FOR USER INTERFACES
A.N. Morozov, V.S. Zarubin, S.A. Grishin
The problem of evaluating the quality of user interfaces of computer programs is considered. Indicators of the overal and relative information content of user interfaces are proposed, their mathematical model is developed, and analytical expressions for calculation are obtained. Most often, an indicator of the quality of user interfaces is understood as a function of the total time of elementary operations on its elements spent by the user to solve a specific task. Interface elements are understood as information interaction objects – input fields, icons, various types of lists, tables, and so on. Basic operations are understood as actions with them – finding, reaching, and managing. The entire set of elementary actions for solving a specific task is called the task profile. It is convenient to use time indicators to get generalized values of the interface quality indicator both in General and at each step of the profile (on the principle that the shorter the time or the number of operations, the better). But they answer the question “how much”, not “why”. Why, for example, can a task profile have a greater number of steps but a shorter execution time compared to a similar one? In General, interface elements make up a complex system with multidimensional relationships by belonging to a particular task, by importance, by frequency and method of use, and so on. Therefore, an attempt to randomly change the configuration of the monitor screen in order to optimize actions at a particular step may lead to changes in the time values at the remaining steps of the profile with unpredictable results. One of the ways to solve the problem of analyzing and/or optimizing interfaces is to develop quality indicators that operate on the characteristics of the overall layout of interfaces, in particular the number, size, and location of elements on the monitor screen. These indicators include the proposed indicators of overall and relative information content.
Keywords: computer program, user interface, analysis, quality indicator, indicator of overall information content, indicator of relative information content.
ONE OF THE ASPECTS OF DEVELOPING A UNIVERSAL TASK DESIGNER FOR TRAINING SYSTEMS
V.V. Salnikov, M.N. Kravtsov
The paper considers the issue of building a block hierarchy of software and information models that simulate the subject area in training systems, based on the principle from simple to complex, from elementary functions to more complex ones. This allows us to implement the principle of adaptability in training systems – the ability to manage the complexity of the training process depending on the competence of the trainee. An approach to the construction of a universal task constructor for training systems based on the representation of the training process in the form of changing a finite number of parameters of the training block by affecting it by the trainee is proposed. The training block in this setting is considered as a set of parameters of various types. A method for evaluating the training process is proposed, which involves checking the state of the block at the end of training, that is, checking whether the parameter values match the specified ones. It allows you to create flexible systems for evaluating knowledge. Correlation of certain parameters with the skills of the trainee allows automating the process of controlling the level of his subject competence and makes it possible to implement the principle of adaptability of training systems, which consists in a gradual increase in the complexity of the training process. A prototype of a universal task designer for training systems is presented. Its use significantly simplifies task editing, neglecting the use of programming environments.
Keywords: training, training, training process, training systems, simulator, task modeling.
STAFF PERFORMANCE MANAGEMENT IN THE CONTEXT OF DIGITAL TRANSFORMATION OF ORGANIZATIONAL SYSTEMS
The article examines the characteristic changes in the staff activity in the conditions of digital transformation of organizational systems. It is shown that the changes are primarily related to the increasing role of the enterprise information system, which provides all types of interaction between the structural components of a new type of organization. With the functioning of the information system itself as a man-machine system, the importance of the effectiveness of staff activities increases. It is proposed to consider the process of staff performance management from the point of view of ensuring its adaptation to new labor functions in the conditions of digital transformation of organizational systems. The components of educational resources for basic and practice-oriented staff training are considered as management tools. The expediency of management decisions based on the optimization approach is proved. A sequence of tasks of reduction, aggregation-balance and resource optimization is formed, which allows you to choose a solution for a variety of thematic modules of preparation for performing labor functions in the conditions of digital transformation, taking into account balance and resource constraints. The resulting solution allows you to solve the task of managing the effectiveness of personnel activities, taking into account the types of activities and labor functions that differ significantly from the traditional ones.
Keywords: management, digital transformation, organizational system, optimization, educational resources.
DISTRIBUTION MODEL OF VOLUMES AND PRICING OF A MATERIAL FLOW IN THE LOGISTIC CHAIN “THE PRODUCER – END USER”
A.S. Dulesov, I.A. Gimanova, O.L. Melnikova, V.I. Yakovchenko
The work considers the construction of an economic and mathematical distribution model of volumes and pricing in logistic channels of the single-market trade-brokerage network. The logistic chain “the producer – the end user” is investigated with successively connected agents through micromarkets. Conditions for a model construction are described. Each participant of a network has its own parameters. Special attention is paid to the coefficient of goods sale for each economic agent and for the whole chain. The problems for three and four participants of a sequential chain with a given price distribution and a uniform amount of product promotion are solved in practice (an ideal case that is not available in practice due to the uncertainty of the information in the form of random influences on the dynamics of the indicators). The solution based on the value of indicators is presented, taking into account the purchases/sales experience, individual preferences and the added price of each agent. On the basis of the obtained values of the sales coefficient of goods (when considering the real situation of the goods promotion), conclusions on the further behavior of participants in a chain are presented. The plan is proposed to adjust the results to the demand of the end user. Volumes and coefficients of realization of the goods with respect to balance between supply and demand are determined. The built-in economic and mathematical distribution model of volumes and pricing will make it possible to develop and make the decision on the choice of a transit or warehouse supply chain.
Keywords: modelling, trade-commerce network, demand and supply, logistic chain, сoefficient of goods sale.
DECISION SUPPORT BASED ON ERROR REPORT CLUSTERING IN COMPLEX OPEN ENDED ASSIGNMENTS QUALITY CONTROL
UDC 004.85, 005.9
There are processes among the processes of different organizations related to carrying out tasks, implementation of which is controlled manually. This is because of a lack of result-template for the tasks. There is only the system of requirements, which implemented task must satisfy. These tasks are known as complex open ended assignments in online learning. However, the tasks exist in other fields, for example, in the publication process, in the equipment and device production process, etc. Complex open ended assignment quality control stage is ineffective due to time-consuming work of an inspector, who checks the conformity of the tasks against the requirements and prepares feedback for a performer. Intellectual support is beginning to be used for a series of tasks. Intellectual support is based upon automatic task implementation classification with the use of machine learning. However, automatic classification can bring to incorrect task implementation quality assessment. Also classifier does not generate a detailed feedback, which fit for a revision of implemented task. A decision support method based on error report clustering, which allows to create a detailed feedback on implemented complex open ended assignments, is suggested in the paper. Special software, which in conjunction with existing clustering system Carrot2 executes suggested method, is developed. The software is introduced in online pre-defense of graduation qualification thesis process. This led to time reduction in feedback preparing by an inspector.
Keywords: data mining, Internet, crime, forecasting, electronic commerce, a posteriori probability.
MODELS AND METHODS FOR SENTIMENT ANALYSIS OF TEXTS IN BASHKIR LANGUAGE
A.K. Suleymanov, M.A. Sharipova, O.N. Smetanina, Y.Y. Sazonova, K.V. Mironov
The research works on automatic opinion extraction are still relevant. The article presents a formal description of the term opinion, setting tasks depending on the determined properties of opinion. The problems of solving the tasks of sentiment analysis, approaches to its solution and ready-made software implementations are described. Available corpora of texts in the Bashkir language are presented, and also task statement for sentiment analysis in the Bashkir language. Presented solution, which include an algorithm for tagging the texts, a preprocessing algorithm, a choice of classification features, and classification algorithms. Also, the results of computational experiment, which aimed to define the most effective classifier based on quality metric, are present. The results in this work and the developed software solution based on SVM with stochastic gradient descent, which demonstrated the highest indicators in the criteria of accuracy, completeness, and 𝐹-measure, can be used to sentiment analysis of news sites in the Bashkir language.
Keywords: sentiment analysis, computational linguistics, machine learning, classification features, hybrid intelligent system, support vector machine, random forest.
THE RESEARCHING OF THE SOCIAL NETWORKS PUBLICATIONS CLASSIFICATION PROBLEM ON THE SUBJECT OF POSITIVE ATTITUDE IDENTIFICATION
M.A. Sazonov, S.V. Shekshuev
In article discusses the relevance of solving problems class publication activity analysis for users of social networks. An analysis of existing approaches identifying public opinion about publications in social networks is given, in which the prevalence is substantiated of methods based on the analysis of the texts sentiment. The disadvantages of these methods are given, which reduce the process of assessing public opinion regarding the publication activity of users of social networks efficiency. It is suggested that it is possible to use message metadata without the need a texts sentiment analysis procedure to eliminate this problem. The primary and derived indicators of messages in social networks are determined, obtained from the set of metadata. Approaches to solving the problem of binary classification based on the indicated markers, both based on statistical methods and using machine learning methods, are considered. An assumption is made about the acceptable accuracy of a class of models based on machine learning that provide a solution to the specified problem. A machine learning model based on a random forest is proposed for solving the problem of classifying a positive attitude towards publications in social networks, based on the analysis of primary and derived indicators of messages.
Keywords: social network, data, social networks publications indicators, machine learning, random forest.
MODEL FOR MATCHING PORTRAITS OF TEACHERS AND TRAINEES
A.V. Ganicheva, A.V. Ganichev
The relevance of this work is due to the need to take into account the personal qualities of teachers and students in the organization of the educational process. The importance of solving this problem is determined by the possibility of forming individual learning paths for students. A set of characteristic features of the teacher forms his professional psychological and pedagogical portrait. The totality of the student’s personality traits forms his psychological portrait. A joint portrait of the teacher and student shows the qualities of both and characterizes their joint activity in the educational process. To describe the portraits, the article developed a mathematical model based on the generating automaton grammar. The principles of building a model of a dynamic portrait are formulated. The characteristics of portraits are defined: volume, weight, significance, importance, mass. For each characteristic, the speed of its change over time is considered. The concepts of activity and efforts to change the activity of the portrait, the complexity of the portrait, and the vector space of the complexity of portraits are introduced. As an effective characteristic of a portrait, work on changing the activity of a portrait and changing the mass of a portrait is considered. Graphical representations of portraits use graphs of a tree structure. To illustrate the results, a specific example is considered. The results of his decision are presented graphically. For the portrait represented by the graph of the tree structure, the characteristics of connectivity, depth, complexity, quality are used.
Keywords:quality of teachers and students, influence coefficient, characteristics of portraits, generative grammar, fragment, portrait tree, balance condition.
A GOAL-ORIENTED CHATBOT BASED ON MACHINE LEARNING
T.M.T Nguyen, M.V. Shcherbakov
Nowadays chatbots are becoming very popular in many areas, such as business, banking, healthcare, study, travel tips, etc. The popularity of messaging platforms such as Telegram, Messenger, Whatsapp, and others has made chatbots not only popular but also become a trend in the future. Since the end of December 2019, the onset of the COVID-19 pandemic has brought about a major global health crisis. Therefore, it is extremely important to provide information about the epidemic to all people. Many governments and organizations have launched chatbots to inform the public about COVID-19. However, these chat rules are limited as they understand a limited set of questions entered by users. Thereby, creating a chatbot based on machine learning for coronavirus information is an urgent task. The purpose of the study is the development of a chatbot for searching for information about COVID-19 coronavirus infection. The method of designing and developing a chatbot on the RASA framework, as well as testing of the developed prototype, are described. Three chatbot models were created: the baseline model (B), the baseline model with synonyms (BS), and the baseline model with synonyms and noises (BSS). The effectiveness of three models was evaluated based on the following indicators: accuracy, precision, and F-measure. The analysis results showed that the BS and BSS models are better than the B model.
Keywords:chat bot, natural language processing, serverless, intent, entities, RASA, COVID-19.