Dmytro I. Uhryn, Artem O. Karachevtsev, Serhii F. Shevchuk, Andrii D. Uhryn
Stock market valuation uses a variety of indicators, such as indices and ratings, to reflect its state and movement. For example, a stock exchange index reflects activity on a stock exchange and is calculated using specific formulas. The calculation of indices is based on statistical data on securities and helps to assess the risks of investments. These indices reflect market conditions. The methodology for forming stock indices includes four stages: sampling, weighting of shares, calculation of the average, and conversion to the index form. Two types of sampling are used: deterministic and floating-power sampling. The weighting coefficients are determined by the price criterion and market capitalization. The studied approaches to stock market modeling allow identifying functional dependencies in the data and developing forecasts. In particular, the methods of approximation and modeling by the Wiener process are allocated. Stock market forecasting using the multi-layer architecture of Long Short-Term Memory in the Keras library is investigated. The overall results confirm that an intelligent information system for automated trading decisions is effective, providing traders with competitive advantages and reducing risks.
股市估值使用指数和评级等各种指标来反映其状态和走势。例如，证券交易所指数反映证券交易所的活动，使用特定公式计算。指数的计算以证券统计数据为基础，有助于评估投资风险。这些指数反映了市场状况。股票指数的编制方法包括四个阶段：取样、股票加权、计算平均值和转换成指数形式。抽样分为两种：确定性抽样和浮动权抽样。加权系数由价格标准和市值决定。所研究的股票市场建模方法可以识别数据中的函数依赖关系并进行预测。特别是分配了维纳过程的近似和建模方法。使用 Keras 库中的长短期记忆多层架构对股市预测进行了研究。总体结果证实，用于自动交易决策的智能信息系统是有效的，可为交易者提供竞争优势并降低风险。
{"title":"Modeling and forecasting of stock market processes","authors":"Dmytro I. Uhryn, Artem O. Karachevtsev, Serhii F. Shevchuk, Andrii D. Uhryn","doi":"10.15276/hait.07.2024.7","DOIUrl":"https://doi.org/10.15276/hait.07.2024.7","url":null,"abstract":"Stock market valuation uses a variety of indicators, such as indices and ratings, to reflect its state and movement. For example, a stock exchange index reflects activity on a stock exchange and is calculated using specific formulas. The calculation of indices is based on statistical data on securities and helps to assess the risks of investments. These indices reflect market conditions. The methodology for forming stock indices includes four stages: sampling, weighting of shares, calculation of the average, and conversion to the index form. Two types of sampling are used: deterministic and floating-power sampling. The weighting coefficients are determined by the price criterion and market capitalization. The studied approaches to stock market modeling allow identifying functional dependencies in the data and developing forecasts. In particular, the methods of approximation and modeling by the Wiener process are allocated. Stock market forecasting using the multi-layer architecture of Long Short-Term Memory in the Keras library is investigated. The overall results confirm that an intelligent information system for automated trading decisions is effective, providing traders with competitive advantages and reducing risks.","PeriodicalId":375628,"journal":{"name":"Herald of Advanced Information Technology","volume":"100 4","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-04-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140747040","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The paper is devoted to the problem of assessing the quality of machine learning models in the form of neural networks in the presence of several requirements for the quality of intelligent systems. The aim of this paper is to develop a multifactorial information criterion that allows choosing a machine learning model in the form of a neural network that best meets the set of requirements for accuracy and interpretability. This goal is achieved through the development and adaptation of multifactorial information criteria for evaluating models in the form of neural networks and, in a particular case, three-layer time delay neural networks used to identify nonlinear dynamic objects. The scientific novelty of the work lies in the development of multifactorial information criteria for the quality of machine learning models that take into account the accuracy and complexity indicators, which, unlike existing information criteria, are adapted to the evaluation of models in the form of neural networks. The practical usefulness of the work lies in the possibility of automatic selection of the simplest machine learning model that provides suitable accuracy when used in intelligent systems. The practical significance of the obtained results lies in the application of the proposed criteria for selecting a machine learning model in the form of a time delay neural network for identifying nonlinear dynamic objects, which allows to increase the accuracy of modeling while ensuring the simplest architecture of the neural network.
{"title":"Assessment of the quality of neural network models based on a multifactorial information criterion","authors":"Oleksandr O. Fomin, V.A. Krykun","doi":"10.15276/hait.07.2024.1","DOIUrl":"https://doi.org/10.15276/hait.07.2024.1","url":null,"abstract":"The paper is devoted to the problem of assessing the quality of machine learning models in the form of neural networks in the presence of several requirements for the quality of intelligent systems. The aim of this paper is to develop a multifactorial information criterion that allows choosing a machine learning model in the form of a neural network that best meets the set of requirements for accuracy and interpretability. This goal is achieved through the development and adaptation of multifactorial information criteria for evaluating models in the form of neural networks and, in a particular case, three-layer time delay neural networks used to identify nonlinear dynamic objects. The scientific novelty of the work lies in the development of multifactorial information criteria for the quality of machine learning models that take into account the accuracy and complexity indicators, which, unlike existing information criteria, are adapted to the evaluation of models in the form of neural networks. The practical usefulness of the work lies in the possibility of automatic selection of the simplest machine learning model that provides suitable accuracy when used in intelligent systems. The practical significance of the obtained results lies in the application of the proposed criteria for selecting a machine learning model in the form of a time delay neural network for identifying nonlinear dynamic objects, which allows to increase the accuracy of modeling while ensuring the simplest architecture of the neural network.","PeriodicalId":375628,"journal":{"name":"Herald of Advanced Information Technology","volume":"70 3","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-04-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140747424","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
A.V. Timenko, V. Shkarupylo, N.A. Kulykovska, Svitlana S. Hrushko
Amid the rapid development of the Internet of Things and its impact on various areas of life, ensuring compatibility between different system components is becoming an urgent task. This is especially important in the context of developing and integrating Internet of Things systems with a high level of diversity and dynamism. In this article, we consider the problem of interoperability of Internet of Things components, focusing on application layer protocols that are key to ensuring intercomponent interaction. The main purpose of the article is to develop and validate a model that will optimize the processes of interaction between system components, taking into account the specifics of protocols. The model is based on the use of temporal action logic, which provides formal verification of interactions between components and allows identifying potential compatibility problems at the early stages of development. The developed model has been tested using a software simulator that allows simulating various scenarios of interaction in the Internet of Things network. The experimental results demonstrate the effectiveness of the proposed methodology in increasing the level of interoperability between system components, which in turn reduces the risks of data loss and ensures the stability of Internet of Things systems. Due to the in-depth analysis and development of specialized methods and tools, this study makes a significant contribution to the development of theoretical and practical aspects of interoperability. However, to further improve the accuracy and versatility of the model, additional empirical studies with a larger data set are recommended.
{"title":"Study of the method of controlling the compatibility of Internet of Things devices based on the MQTT application layer protocol","authors":"A.V. Timenko, V. Shkarupylo, N.A. Kulykovska, Svitlana S. Hrushko","doi":"10.15276/hait.07.2024.4","DOIUrl":"https://doi.org/10.15276/hait.07.2024.4","url":null,"abstract":"Amid the rapid development of the Internet of Things and its impact on various areas of life, ensuring compatibility between different system components is becoming an urgent task. This is especially important in the context of developing and integrating Internet of Things systems with a high level of diversity and dynamism. In this article, we consider the problem of interoperability of Internet of Things components, focusing on application layer protocols that are key to ensuring intercomponent interaction. The main purpose of the article is to develop and validate a model that will optimize the processes of interaction between system components, taking into account the specifics of protocols. The model is based on the use of temporal action logic, which provides formal verification of interactions between components and allows identifying potential compatibility problems at the early stages of development. The developed model has been tested using a software simulator that allows simulating various scenarios of interaction in the Internet of Things network. The experimental results demonstrate the effectiveness of the proposed methodology in increasing the level of interoperability between system components, which in turn reduces the risks of data loss and ensures the stability of Internet of Things systems. Due to the in-depth analysis and development of specialized methods and tools, this study makes a significant contribution to the development of theoretical and practical aspects of interoperability. However, to further improve the accuracy and versatility of the model, additional empirical studies with a larger data set are recommended.","PeriodicalId":375628,"journal":{"name":"Herald of Advanced Information Technology","volume":"95 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-04-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140750547","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Augmented reality is an innovative technology that merges the virtual and real worlds to create a unique interactive experience for users. Although the technology only became widely known in the mid-2000s, its potential and applications continue to evolve rapidly, especially with the advent of smartphones and other mobile devices that allow a wide range of users to interact with augmented reality in their everyday lives. In the field of cultural heritage and tourism, augmented reality opens up new opportunities for the restoration and presentation of historical buildings and places that no longer exist or have been altered by time, giving visitors the opportunity to see and experience the historical environment in its original form. However, to effectively use augmented reality in this area, it is necessary to ensure an exact correspondence between virtual objects and the real environment, as well as to implement navigation functions that will help users easily navigate the virtually restored space. Developing and implementing augmented reality solutions requires not only technological expertise but also a deep understanding of the historical, cultural and social context of the objects being recreated. Based on our research, we have developed a concept of an augmented reality application for the reconstruction and promotion of cultural heritage. This includes methods for accurately recreating historical locations and objects in a virtual environment, as well as developing intuitive navigation tools for users. The main achievement of the work is the creation of a foundation for the further development of augmented reality technologies in this area, with a focus on improving the interaction between virtual and real components, which will help increase audience engagement and raise awareness of historical and cultural heritage. This opens up broad prospects for the use of augmented reality for cultural heritage, and the proposed approaches can serve as a basis for future innovative projects in this area.
{"title":"The use of augmented reality for renovation of cultural heritage sites","authors":"A. Sachenko, Ivan R. Kit","doi":"10.15276/hait.07.2024.2","DOIUrl":"https://doi.org/10.15276/hait.07.2024.2","url":null,"abstract":"Augmented reality is an innovative technology that merges the virtual and real worlds to create a unique interactive experience for users. Although the technology only became widely known in the mid-2000s, its potential and applications continue to evolve rapidly, especially with the advent of smartphones and other mobile devices that allow a wide range of users to interact with augmented reality in their everyday lives. In the field of cultural heritage and tourism, augmented reality opens up new opportunities for the restoration and presentation of historical buildings and places that no longer exist or have been altered by time, giving visitors the opportunity to see and experience the historical environment in its original form. However, to effectively use augmented reality in this area, it is necessary to ensure an exact correspondence between virtual objects and the real environment, as well as to implement navigation functions that will help users easily navigate the virtually restored space. Developing and implementing augmented reality solutions requires not only technological expertise but also a deep understanding of the historical, cultural and social context of the objects being recreated. Based on our research, we have developed a concept of an augmented reality application for the reconstruction and promotion of cultural heritage. This includes methods for accurately recreating historical locations and objects in a virtual environment, as well as developing intuitive navigation tools for users. The main achievement of the work is the creation of a foundation for the further development of augmented reality technologies in this area, with a focus on improving the interaction between virtual and real components, which will help increase audience engagement and raise awareness of historical and cultural heritage. This opens up broad prospects for the use of augmented reality for cultural heritage, and the proposed approaches can serve as a basis for future innovative projects in this area.","PeriodicalId":375628,"journal":{"name":"Herald of Advanced Information Technology","volume":"47 5","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-04-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140747296","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Conventional group recommender systems fail to take into account the impact of group dynamics on group recommendations, such as the process of reconciling individual preferences during collective decision-making. This scenario has been previously examined in the context of group decision making, specifically in relation to consensus reaching procedures. In such processes, experts engage in negotiations to determine their preferences and ultimately pick a mutually agreed upon option. The objective of the consensus procedure is to prevent dissatisfaction among group members about the suggestion. Prior studies have tried to accomplish this characteristic in group recommendation by using the minimal operator for the process of aggregating recommendations. Nevertheless, the use of this operator ensures just a minimal degree of consensus on the proposal, but it does not provide a satisfactory level of agreement among group members over the group recommendation. This paper focuses on analyzing consensus reaching procedures in the context of group recommendation for group decision making. The goal of the study is to use consensus reaching processes to provide group recommendations that satisfy all members of the group. Additionally, study aims to enhance group recommender systems by ensuring an acceptable level of agreement among users regarding the group recommendation. Therefore, group recommender systems are expanded by including consensus reaching mechanisms to facilitate group decision making. In the context of group decision making, a collective resolution is reached by a group of persons, who may be specialists, from a pool of options or potential solutions to the issue at hand. To do this, each specialist obtains their preferences about each possibility. The conventional selection techniques for group decision-making difficulties fail to include the possibility of dissent among experts over the chosen choice. This issue is alleviated by using consensus-building techniques, in which a substantial degree of agreement is attained prior to picking the ultimate decision. To facilitate alignment of experts' tastes, they repeatedly modify them to increase their proximity. Prior to making collective choices, it is sometimes necessary to establish a certain degree of consensus. Thus, this paper presents a group recommendation architecture that utilizes automated consensus reaching models to provide accepted suggestions. More specifically, we are considering the minimal cost consensus model and the automated consensus support system model that relies on input. The minimal cost consensus model calculates the collective suggestion of a group by adjusting individual preferences based on a cost function. This is achieved via the use of linear programming. The feedback-based automated consensus support system model mimics the interaction between group members and a moderator. The moderator offers adjustments to individual suggestions in order to bring them clos
{"title":"Reaching consensus in group recommendation systems","authors":"Anastasiia A. Gorbatenko, Mykola Hodovychenko","doi":"10.15276/hait.07.2024.3","DOIUrl":"https://doi.org/10.15276/hait.07.2024.3","url":null,"abstract":"Conventional group recommender systems fail to take into account the impact of group dynamics on group recommendations, such as the process of reconciling individual preferences during collective decision-making. This scenario has been previously examined in the context of group decision making, specifically in relation to consensus reaching procedures. In such processes, experts engage in negotiations to determine their preferences and ultimately pick a mutually agreed upon option. The objective of the consensus procedure is to prevent dissatisfaction among group members about the suggestion. Prior studies have tried to accomplish this characteristic in group recommendation by using the minimal operator for the process of aggregating recommendations. Nevertheless, the use of this operator ensures just a minimal degree of consensus on the proposal, but it does not provide a satisfactory level of agreement among group members over the group recommendation. This paper focuses on analyzing consensus reaching procedures in the context of group recommendation for group decision making. The goal of the study is to use consensus reaching processes to provide group recommendations that satisfy all members of the group. Additionally, study aims to enhance group recommender systems by ensuring an acceptable level of agreement among users regarding the group recommendation. Therefore, group recommender systems are expanded by including consensus reaching mechanisms to facilitate group decision making. In the context of group decision making, a collective resolution is reached by a group of persons, who may be specialists, from a pool of options or potential solutions to the issue at hand. To do this, each specialist obtains their preferences about each possibility. The conventional selection techniques for group decision-making difficulties fail to include the possibility of dissent among experts over the chosen choice. This issue is alleviated by using consensus-building techniques, in which a substantial degree of agreement is attained prior to picking the ultimate decision. To facilitate alignment of experts' tastes, they repeatedly modify them to increase their proximity. Prior to making collective choices, it is sometimes necessary to establish a certain degree of consensus. Thus, this paper presents a group recommendation architecture that utilizes automated consensus reaching models to provide accepted suggestions. More specifically, we are considering the minimal cost consensus model and the automated consensus support system model that relies on input. The minimal cost consensus model calculates the collective suggestion of a group by adjusting individual preferences based on a cost function. This is achieved via the use of linear programming. The feedback-based automated consensus support system model mimics the interaction between group members and a moderator. The moderator offers adjustments to individual suggestions in order to bring them clos","PeriodicalId":375628,"journal":{"name":"Herald of Advanced Information Technology","volume":"12 2","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-04-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140747008","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Currently, light electric vehicles are rapidly developing in various kinds. To power these vehicles with batteries, the simplest electric drive system is a DC motor controlled by a DC-DC converter. This work utilizes a bidirectional Zeta-SEPIC DC-DC converter with an integrated DC motor. This implementation enables control of motor speed and torque in traction and regenerative braking modes. Additionally, it allows for the use of a lower voltage battery compared to the motor's rated voltage, reducing battery weight and increasing safety. In this work, a decomposition approach is applied. Two separate port-controlled Hamiltonian subsystems are obtained to adjust the motor angular velocity in the traction (Zeta) and braking (SEPIC) modes of the DC-DC converter. The Passivity-Based Control (PBC) method is used to synthesize the drive control subsystems in these modes. This method is based on the energy laws of processes in systems and provides asymptotic stability of nonlinear systems, in this case, two fourth-order subsystems for speed control. Two third-order current control subsystems synthesized by the PBC were used to limit the motor current at a given level. The synthesis resulted in sets of possible structures of control influence formers (CIFs) for all PBC subsystems using Zeta and SEPIC DC-DC converters. The study analyzed the operation of the obtained structures of the CIFs, selected the most effective ones, and determined the laws of adaptation of their parameters to the value of the motor angular velocity through computer simulation in Matlab/Simulink. The results of the simulation showed that the drive operated well in both static and dynamic modes.
{"title":"Mode decomposed passivity-based speed control of DC drive with bidirectional Zeta-SEPIC DC-DC converter for light electric vehicles","authors":"Rostyslav-Ivan Kuzyk, I. Shchur","doi":"10.15276/hait.07.2024.6","DOIUrl":"https://doi.org/10.15276/hait.07.2024.6","url":null,"abstract":"Currently, light electric vehicles are rapidly developing in various kinds. To power these vehicles with batteries, the simplest electric drive system is a DC motor controlled by a DC-DC converter. This work utilizes a bidirectional Zeta-SEPIC DC-DC converter with an integrated DC motor. This implementation enables control of motor speed and torque in traction and regenerative braking modes. Additionally, it allows for the use of a lower voltage battery compared to the motor's rated voltage, reducing battery weight and increasing safety. In this work, a decomposition approach is applied. Two separate port-controlled Hamiltonian subsystems are obtained to adjust the motor angular velocity in the traction (Zeta) and braking (SEPIC) modes of the DC-DC converter. The Passivity-Based Control (PBC) method is used to synthesize the drive control subsystems in these modes. This method is based on the energy laws of processes in systems and provides asymptotic stability of nonlinear systems, in this case, two fourth-order subsystems for speed control. Two third-order current control subsystems synthesized by the PBC were used to limit the motor current at a given level. The synthesis resulted in sets of possible structures of control influence formers (CIFs) for all PBC subsystems using Zeta and SEPIC DC-DC converters. The study analyzed the operation of the obtained structures of the CIFs, selected the most effective ones, and determined the laws of adaptation of their parameters to the value of the motor angular velocity through computer simulation in Matlab/Simulink. The results of the simulation showed that the drive operated well in both static and dynamic modes.","PeriodicalId":375628,"journal":{"name":"Herald of Advanced Information Technology","volume":"580 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-04-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140749407","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Vladimir P. Zaykov, V. Mescheryakov, Andrey S. Ustenko
The paper presents the results of research of controllability of the thermoelectric system for ensuring thermal modes of electronic equipment, including a regulator, a cooler, and a component of excess heat removal to the environment. It is shown that for the use of methods of optimization of automatic control systems it is necessary to study the transfer and dynamic characteristics of the object - thermoelectric cooling device with one input and one output. The mathematical model of the thermoelectric cooler of the system of providing thermal modes of a given design is presented, which takes into account the influence of the conditions of interaction of the heat sink with the medium on the main significant parameters, reliability indicators, dynamic and energy characteristics of the single stage cooler. The model is created for the operating range of cooling, level of thermal load, geometry of thermocouple branches, different temperature of the medium, for the characteristic thermal regime of maximum cooling capacity. The results of calculations of the main significant parameters, reliability indicators, dynamic and energy characteristics of the cooler for different medium temperature in the operating temperature range and variation of conditions of heat exchange of the heat sink with the medium are given. It is shown that as the intensity of heat exchange of the heat sink with the medium increases, the temperature difference between the heat sink and the medium decreases. This makes it possible to significantly reduce the relative failure rate, increase the probability of failure-free operation of the thermoelectric cooler and control the reliability indicators of the device of a given design during operation.
{"title":"Method of reliability control of thermoelectric systems to ensure thermal regimes","authors":"Vladimir P. Zaykov, V. Mescheryakov, Andrey S. Ustenko","doi":"10.15276/hait.07.2024.5","DOIUrl":"https://doi.org/10.15276/hait.07.2024.5","url":null,"abstract":"The paper presents the results of research of controllability of the thermoelectric system for ensuring thermal modes of electronic equipment, including a regulator, a cooler, and a component of excess heat removal to the environment. It is shown that for the use of methods of optimization of automatic control systems it is necessary to study the transfer and dynamic characteristics of the object - thermoelectric cooling device with one input and one output. The mathematical model of the thermoelectric cooler of the system of providing thermal modes of a given design is presented, which takes into account the influence of the conditions of interaction of the heat sink with the medium on the main significant parameters, reliability indicators, dynamic and energy characteristics of the single stage cooler. The model is created for the operating range of cooling, level of thermal load, geometry of thermocouple branches, different temperature of the medium, for the characteristic thermal regime of maximum cooling capacity. The results of calculations of the main significant parameters, reliability indicators, dynamic and energy characteristics of the cooler for different medium temperature in the operating temperature range and variation of conditions of heat exchange of the heat sink with the medium are given. It is shown that as the intensity of heat exchange of the heat sink with the medium increases, the temperature difference between the heat sink and the medium decreases. This makes it possible to significantly reduce the relative failure rate, increase the probability of failure-free operation of the thermoelectric cooler and control the reliability indicators of the device of a given design during operation.","PeriodicalId":375628,"journal":{"name":"Herald of Advanced Information Technology","volume":"189 3","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-04-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140746471","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-10-12DOI: 10.15276/hait.06.2023.13
Olena О. Arsirii, Denys V. Petrosiuk
The relevance of solving the problem of facial emotion recognition on human images in the creation of modern intelligent systems of computer vision and human-machine interaction, online learning and emotional marketing, health care and forensics, machine graphics and game intelligence is shown. Successful examples of technological solutions to the problem of facial emotion recognition using transfer learning of deep convolutional neural networks are shown. But the use of such popular datasets as DISFA, CelebA, AffectNet, for deep learning of convolutional neuralnetworks does not give good results in terms of the accuracy of emotion recognition, because almost all training sets have fundamental flaws related to errors in their creation, such as the lack of data of a certain class, imbalance of classes, subjectivity and ambiguity of labeling, insufficient amount of data for deep learning, etc. It is proposed to overcome the noted shortcomings of popular datasets for emotion recognition by adding to the training sample additional pseudo-labeled images with human emotions, on which recognition occurs with high accuracy. The aim of the research is to increase the accuracy of facial emotion recognitionon the image of a human by developing a pseudo-labeling method for transfer learning of a deep neural network. To achieve the aim, the following tasks were solved: a convolutional neural network model, previously trained on the ImageNet set using the transfer learning method, was adjusted on the RAF-DB data set to solve emotion recognition tasks; a pseudo-labeling method of the RAF−DB set data was developed for semi-supervised learning of a convolutional neural network model for the task of facial emotion recognition; the accuracy of facial emotion recognition was analyzed based on the developed convolutional neural network model and the method of pseudo-labeling of RAF-DB set data for its correction. It is shown that the use of the developed method of pseudo-labeling data and transfer learning of the MobileNet V1 convolutional neural network model allowed to increase the accuracy of facial emotion recognitionon the images of the RAF-DB dataset by 2 percent (from 76 to 78%) according to the F1 estimate. Atthe same time, taking into account the significant imbalance of the classes, for the 7 main emotions in the trainingset, we have a significant increase in the accuracy of recognizing a few representatives of such emotions as surprise (from 71 to 77%), fearful(from 64 to 69%), sad (from 72 to 76%), angrywith (from 64 to 74%), neutral(from 66 to 71%). The accuracy of recognizing the emotion of happy, which is the most common, decreased (from 91 to 86 %) Thus, it can be concluded that the use of the developed pseudo-labeling method gives good results in overcoming such shortcomings of datasets for deep learning of convolutional neural networks such as lack of data of a certain type, imbalance of classes, insufficient amount of data for deep learning, etc.
{"title":"Pseudo-labeling of transfer learning convolutional neural network data for human facial emotion recognition","authors":"Olena О. Arsirii, Denys V. Petrosiuk","doi":"10.15276/hait.06.2023.13","DOIUrl":"https://doi.org/10.15276/hait.06.2023.13","url":null,"abstract":"The relevance of solving the problem of facial emotion recognition on human images in the creation of modern intelligent systems of computer vision and human-machine interaction, online learning and emotional marketing, health care and forensics, machine graphics and game intelligence is shown. Successful examples of technological solutions to the problem of facial emotion recognition using transfer learning of deep convolutional neural networks are shown. But the use of such popular datasets as DISFA, CelebA, AffectNet, for deep learning of convolutional neuralnetworks does not give good results in terms of the accuracy of emotion recognition, because almost all training sets have fundamental flaws related to errors in their creation, such as the lack of data of a certain class, imbalance of classes, subjectivity and ambiguity of labeling, insufficient amount of data for deep learning, etc. It is proposed to overcome the noted shortcomings of popular datasets for emotion recognition by adding to the training sample additional pseudo-labeled images with human emotions, on which recognition occurs with high accuracy. The aim of the research is to increase the accuracy of facial emotion recognitionon the image of a human by developing a pseudo-labeling method for transfer learning of a deep neural network. To achieve the aim, the following tasks were solved: a convolutional neural network model, previously trained on the ImageNet set using the transfer learning method, was adjusted on the RAF-DB data set to solve emotion recognition tasks; a pseudo-labeling method of the RAF−DB set data was developed for semi-supervised learning of a convolutional neural network model for the task of facial emotion recognition; the accuracy of facial emotion recognition was analyzed based on the developed convolutional neural network model and the method of pseudo-labeling of RAF-DB set data for its correction. It is shown that the use of the developed method of pseudo-labeling data and transfer learning of the MobileNet V1 convolutional neural network model allowed to increase the accuracy of facial emotion recognitionon the images of the RAF-DB dataset by 2 percent (from 76 to 78%) according to the F1 estimate. Atthe same time, taking into account the significant imbalance of the classes, for the 7 main emotions in the trainingset, we have a significant increase in the accuracy of recognizing a few representatives of such emotions as surprise (from 71 to 77%), fearful(from 64 to 69%), sad (from 72 to 76%), angrywith (from 64 to 74%), neutral(from 66 to 71%). The accuracy of recognizing the emotion of happy, which is the most common, decreased (from 91 to 86 %) Thus, it can be concluded that the use of the developed pseudo-labeling method gives good results in overcoming such shortcomings of datasets for deep learning of convolutional neural networks such as lack of data of a certain type, imbalance of classes, insufficient amount of data for deep learning, etc.","PeriodicalId":375628,"journal":{"name":"Herald of Advanced Information Technology","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-10-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136057422","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-10-12DOI: 10.15276/hait.06.2023.18
Mykhaylo V. Lobachev, Sergiy V. Purish
The paper explores the challenge of human identification through gait recognition within biometric identification systems. Itoutlines the essential criteria for human biometric features, discusses primary biometric characteristics, and their application in biometric identification systems. The paper also examines the feasibility of utilizing gait as a biometric identifier, emphasizing its advantages, such as not requiring the upfront provision of personal biometric information and specialized equipment.The authors conduct an analysis of existing scientific literature in the field of gait recognition, categorizing gait recognition methodsinto template-based and non-template-based approaches. Throughout their research, they identify the key issues and challenges that researchers face in this domain, along with the prevailing trends in human gait recognition within biometric identification systems.Additionally, the paper introduces a method for person identification based on gait, utilizing the Histogram of Oriented Gradients and the Sum Variance Haralick texture features. It involves transforming input video into a series of images depicting the gait silhouette, creating a Gait Energy Image (GEI) by combining these gait silhouettes throughout a gait cycle, and translating the GEI into the Gait Gradient Magnitude Image (GGMI). The subsequent step involves extracting recommended gait characteristics from the GGMIs of participants included in a dataset.To preprocess the collected characteristics, Principal Component Analysis (PCA) is applied, reducing the dimensions that may negatively impact classification robustness, thereby enhancing overall performance. In the final step, a K-Nearest Neighbors (KNN) classifier is employed to categorize the characteristics obtained from a specific dataset.The proposed novel feature vector in the paper demonstrates increased reliability and effectively captures spatial variations in gait patterns. Notably, it reduces the dimensionality of the feature vector from 3780×1 to 63×1, resulting in decreased computational complexity in the gait recognition system. Experimental evaluations on the CASIA A and CASIA B datasets reveal that the proposed approach outperforms other HOG-based methods in most scenarios, with the exception of situations involving frontal images.
{"title":"Machine learning models and methods for human gait recognition","authors":"Mykhaylo V. Lobachev, Sergiy V. Purish","doi":"10.15276/hait.06.2023.18","DOIUrl":"https://doi.org/10.15276/hait.06.2023.18","url":null,"abstract":"The paper explores the challenge of human identification through gait recognition within biometric identification systems. Itoutlines the essential criteria for human biometric features, discusses primary biometric characteristics, and their application in biometric identification systems. The paper also examines the feasibility of utilizing gait as a biometric identifier, emphasizing its advantages, such as not requiring the upfront provision of personal biometric information and specialized equipment.The authors conduct an analysis of existing scientific literature in the field of gait recognition, categorizing gait recognition methodsinto template-based and non-template-based approaches. Throughout their research, they identify the key issues and challenges that researchers face in this domain, along with the prevailing trends in human gait recognition within biometric identification systems.Additionally, the paper introduces a method for person identification based on gait, utilizing the Histogram of Oriented Gradients and the Sum Variance Haralick texture features. It involves transforming input video into a series of images depicting the gait silhouette, creating a Gait Energy Image (GEI) by combining these gait silhouettes throughout a gait cycle, and translating the GEI into the Gait Gradient Magnitude Image (GGMI). The subsequent step involves extracting recommended gait characteristics from the GGMIs of participants included in a dataset.To preprocess the collected characteristics, Principal Component Analysis (PCA) is applied, reducing the dimensions that may negatively impact classification robustness, thereby enhancing overall performance. In the final step, a K-Nearest Neighbors (KNN) classifier is employed to categorize the characteristics obtained from a specific dataset.The proposed novel feature vector in the paper demonstrates increased reliability and effectively captures spatial variations in gait patterns. Notably, it reduces the dimensionality of the feature vector from 3780×1 to 63×1, resulting in decreased computational complexity in the gait recognition system. Experimental evaluations on the CASIA A and CASIA B datasets reveal that the proposed approach outperforms other HOG-based methods in most scenarios, with the exception of situations involving frontal images.","PeriodicalId":375628,"journal":{"name":"Herald of Advanced Information Technology","volume":"150 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-10-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136058501","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-10-12DOI: 10.15276/hait.06.2023.17
Anatoly V. Usov, Maksym V. Kunitsyn, Yuriy I. Zaychyk
The purpose of the work is to present the concept of an information modelof the thermomechanical process in grinding of products from materials prone to defect formation due to the fact that their surface layer has hereditary defects of structural or technological origin.The products' strength and functionality depend on the inhomogeneity and defectiveness of the structure of the materials from which they are made. Such materials have many different micro defects formed in the surface layer of parts during the technological operations of their production. Reducing number of defectsin the finishing operations of these materials and increasing the operational properties of products made of these materials is an essential national and economical task, the solution of which leads to a significant saving of material resources, labour intensity and cost of manufacturing parts. The currently available information on the thermal processes of diamond abrasive processing is obtained on the assumption of the homogeneity of the materials being polished and needs to consider the presence of defects in the technological heredity of the products. The phenomenological approach in studying the causes of cracking of materials prone to this type of defect does not allow to revealthe mechanism of genesis and development of grinding cracks. The choice of the method of investigation of the mechanism of crack formation is based on micro-research related to inhomogeneities, which are formed in the surface layer of parts during previous technological operations. A mathematical model has been developed that describes thermomechanical processes in the surface layer during grinding of parts made of materials and alloys, taking into account their inhomogeneities, which affect the intensity of the formation of grinding cracks. Calculated dependences between the crack resistance criterion and the main controlling technological parameters were obtained. According to the known characteristics of hereditary defects, the limit values of thermomechanical criteria, which ensure the necessary quality of the surfaces of the processed products, are determined. Based on the obtained criterion ratios, an algorithm for selecting technological possibilities for defect-free processing of products from materials prone to loss of quality of the surface layer of parts was built. A decision support system has been developed to increase the efficiency of the finishing process management.
{"title":"An information model concept for a thermomechanical process in grinding","authors":"Anatoly V. Usov, Maksym V. Kunitsyn, Yuriy I. Zaychyk","doi":"10.15276/hait.06.2023.17","DOIUrl":"https://doi.org/10.15276/hait.06.2023.17","url":null,"abstract":"The purpose of the work is to present the concept of an information modelof the thermomechanical process in grinding of products from materials prone to defect formation due to the fact that their surface layer has hereditary defects of structural or technological origin.The products' strength and functionality depend on the inhomogeneity and defectiveness of the structure of the materials from which they are made. Such materials have many different micro defects formed in the surface layer of parts during the technological operations of their production. Reducing number of defectsin the finishing operations of these materials and increasing the operational properties of products made of these materials is an essential national and economical task, the solution of which leads to a significant saving of material resources, labour intensity and cost of manufacturing parts. The currently available information on the thermal processes of diamond abrasive processing is obtained on the assumption of the homogeneity of the materials being polished and needs to consider the presence of defects in the technological heredity of the products. The phenomenological approach in studying the causes of cracking of materials prone to this type of defect does not allow to revealthe mechanism of genesis and development of grinding cracks. The choice of the method of investigation of the mechanism of crack formation is based on micro-research related to inhomogeneities, which are formed in the surface layer of parts during previous technological operations. A mathematical model has been developed that describes thermomechanical processes in the surface layer during grinding of parts made of materials and alloys, taking into account their inhomogeneities, which affect the intensity of the formation of grinding cracks. Calculated dependences between the crack resistance criterion and the main controlling technological parameters were obtained. According to the known characteristics of hereditary defects, the limit values of thermomechanical criteria, which ensure the necessary quality of the surfaces of the processed products, are determined. Based on the obtained criterion ratios, an algorithm for selecting technological possibilities for defect-free processing of products from materials prone to loss of quality of the surface layer of parts was built. A decision support system has been developed to increase the efficiency of the finishing process management.","PeriodicalId":375628,"journal":{"name":"Herald of Advanced Information Technology","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-10-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136057423","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：