Bulletin of Electrical Engineering and Informatics
Not a member yet
2885 research outputs found
Sort by
A deep Q-learning approach for adaptive cybersecurity threat detection in dynamic networks
Full text linkCybersecurity faces persistent challenges due to the rapid growth and complexity of network-based threats. Conventional rule-based systems and classical machine learning approaches often lack the adaptability required to detect advanced and dynamic attacks in real time. This study introduces a deep Q-learning framework for autonomous threat detection and mitigation within a simulated network environment that reflects realistic traffic, malicious behaviors, and system conditions. The framework incorporates experience replay and target network stabilization to strengthen learning and policy optimization. Evaluation was performed on a synthesized dataset containing benign traffic and multiple attack categories, including distributed denial of service (DDoS), phishing, advanced persistent threats, and malware. The proposed system achieved 96.7% detection accuracy, an F1-score of 0.94, and reduced detection latency to 50 ms. These results surpassed the performance of rule-based methods and traditional classifiers such as support vector machines, random forests, convolutional neural networks, and recurrent neural networks. A central contribution lies in combining dynamic feature selection with reinforcement learning (RL), allowing the agent to adapt to evolving threats and diverse network conditions. The findings demonstrate the potential of deep reinforcement learning (DRL) as a scalable and efficient solution for real-time cybersecurity defense
Modern artificial intelligence technics for unmanned aerial vehicles path planning and control
Full text linkUnmanned aerial vehicles (UAVs) require effective path planning algorithms to navigate through complex environments. This study investigates the application of Deep Q-learning and Dyna Q-learning methods for UAV path planning and incorporates fuzzy logic for enhanced control. Deep Q-learning, a reinforcement learning technique, employs a deep neural network to approximate Q-values, allowing the UAV to improve its path planning capabilities by maximizing cumulative rewards. Conversely, Dyna Q-learning leverages simulated scenarios to update Q-values, refining the UAV’s decision-making process and adaptability to dynamic environments. Additionally, fuzzy logic control is integrated to manage UAV movements along the planned path. This control system uses linguistic variables and fuzzy rules to handle uncertainties and imprecise information, enabling real-time adjustments to speed, altitude, and heading for accurate path following and obstacle avoidance. The research evaluates the effectiveness of these methods individually, with a focus on model-free learning in a gradual training approach, and compares their performance in terms of path planning accuracy, adaptability, and obstacle avoidance. The paper contributes to a deeper understanding of UAV path planning techniques and their practical applications in various scenarios
A novel hybrid SMOTE oversampling approach for balancing class distribution on social media text
Full text linkDepression is a frequent and dangerous medical disorder that has an unhealthy effect on how a person feels, thinks, and acts. Depression is also quite prevalent. Early detection and treatment of depression may avoid painful and perhaps life-threatening symptoms. An imbalance in the data creates several challenges. Consequently, the majority learners will have biases against the class that constitutes the majority and, in extreme situations, may completely dismiss the class that constitutes the minority. For decades, class disparity research has employed traditional machine learning methods. In addressing the challenge of imbalanced data in depression detection, the study aims to balance class distribution using a hybrid approach bidirectional long short-term memory (BI-LSTM) along with synthetic minority over-sampling and Tomek links and synthetic minority over-sampling and edited nearest neighbors’ techniques. This investigation presents a new approach that combines synthetic minority oversampling technique with the Kalman filter to provide an innovative extension. The Kalman-synthetic minority oversampling technique (KSMOTE) approach filters out noisy samples in the final dataset, which consists of both the original data and the artificially created samples by SMOTE. The result was greater accuracy with the BI-LSTM classification scheme compared to the other standard methods for finding depression in both unbalanced and balanced data
Detection and prevention of Man-in-The-Middle attack in cloud computing using Openstack
Full text linkThis paper proposes a new technique designed to prevent and detect address resolution protocol (ARP) spoofing attacks in general, and specifically Man-in-the-Middle (MitM) attacks, within the context of cloud computing. The solution focuses on establishing appropriate flow filtering rules based on parameters such as 'time feature' and internet control message protocol '(ICMP) protocol'. The tests were conducted using the Openstack platform. One of the key benefits of this proposed approach is the improved performance in effectively detecting a significant number of malicious packets. We implemented this solution on the Openstack platform and conducted evaluations to demonstrate its efficacy. The results confirm that our method achieves superior performance in detecting MitM attacks, with a packet detection ratio (PDR) of 60.4%. Moving forward, this work will contribute to protecting cloud environments from a large number of MitM attacks
Impact of usability on continuance usage intention in language learning apps with gamification features
Full text linkWith the increasing popularity of language learning applications, gamification has become one of the approaches frequently integrated to enhance continuance usage intention (CUI). However, how the usability of these gamification features influences the intention to continue usage is not yet fully understood. Through the usability testing and system usability scale (SUS) method, this study evaluates the level of usability of gamification features in language learning applications with a novel research approach that also involves specific analysis regarding usability aspects according to Nielsen, including learnability, efficiency, memorability, errors, and satisfaction towards the CUI, categorized through the SUS statements grouping and then processed using the SPSS application. The study results indicate that the SUS scores show above-average scores for all three applications: Duolingo application at 77.08, Elsa Speak at 70, and Cake Learn at 70.58. Other findings suggest that usability aspects generally significantly influence CUI; however, only the satisfaction variable impacts CUI, which was observed only in Duolingo and Elsa Speak. These findings indicate that the overall usability of gamification features positively impacts CUI using language learning applications, thereby implying the need for continuous development
Global research trends in building-integrated photovoltaics: a bibliometric analysis (1971-2022)
Full text linkThe number of academic publications in the building-integrated photovoltaics (BIPV) field has rapidly grown. Most published articles focus on a specific topic, such as mathematical model, solar architecture design, photovoltaic effect, solar cell, grid-connected, efficiency, performance assessment, economic analysis, optimization, and others with broader focus areas. This work focuses on BIPV research with bibliometric analysis through documents, cited references, authors, affiliations, countries, funding sponsors, sources, words, and conceptual structure based on the Scopus-indexed database between 1971 and 2022. The result shows that BIPV research constantly grows annually with strong collaboration authorship. China is the most relevant country with the top affiliation and funding sponsor to support the BIPV research. The terms conjugated polymers, photovoltaic properties, and organic polymer are identified as niche themes. On the other hand, the terms of conversion efficiency, perovskite, photovoltaic devices, solar cells, efficiency, and photoelectrochemical cell clusters are emerging themes. In the future, BIPV research will move towards microgrids, energy, performance, energy management systems, and energy efficiency issues. The finding will also provide researchers and organizations with a comprehensive understanding of BIPV research areas and new directions for future research
Initial study of general theory of complex systems: physical basis and philosophical understanding
Full text linkThe fundamental difference between neural networks containing and not containing feedback between elements is analyzed. It is shown that in the first of these cases, quantitative relationships describing the functioning of the neural network can be obtained based on an analogy with the theory of noise-resistant codes. In the second case, an analogy with electronic circuits that form memory cells (triggers) is valid. It is shown that feedback between elements of even the simplest neural networks can lead to the appearance of multidimensional hysteresis, when, with the same state of inputs, the system can be in several qualitatively different states, the transition between which can be abrupt. In this case, the state of the neural network outputs depends not only on the current state of its inputs, but also on the path along which this state was formed. The results obtained are used for the philosophical substantiation of a new approach to the interpretation of complex systems of various natures, which are considered analogs of neural networks. According to it, a system that can store and processing information should be considered "complex"
Web system to enhance technical supervision of incidents at the hydrocarbons regulatory institution in Lima–2023
Full text linkIn this article, the implementation of a web system was carried out to improve the process of technical supervision of incidents of a hydrocarbon regulatory company because time was lost in carrying out each process; this research was developed using the SCRUM methodology as it is an agile methodology and adapted to our research. Using the process, events and artifact, it was possible to design the prototypes of the system, architecture, and database. Finally, the implementation was carried out among other important points obtained as results; the average level of optimization of the incident assignment process, derived from the observations, is 91.05% efficiency in assigning incidents to specialists. Regarding the 95% confidence interval for this indicator, it is between 88.98% and 93.11% efficiency, representing two standard deviations with respect to the mean. Regarding the average response time to incidents in all states, obtained from observations, it is 15 days. The 95% confidence interval for this indicator ranges between 14 and 18 days, which represents two standard deviations from the mean. The system is intuitive and not complex. With the implementation of the web system, processes are automated and end user satisfaction is obtained
Fuzzy logic method for making push notifications on monitoring system of IoT-based electric truck charging
Full text linkTo minimize the negligence when charging electric vehicles, it is deemed important to have an internet of things (IoT) based monitoring system using a notification feature. The monitoring system of electric vehicle battery charging used a voltage divider and temperature sensor (DS18B20) installed on the Arduino Mega 2560 microcontroller with the addition of an ESP8266 Wi-Fi module for sending microcontroller data into the Blynk platform. A notification feature was added as the reminder that the battery has been overcharging or overheating. This study applied the Mamdani fuzzy logic method to determine the conditions when notifications must appear. The results of the application of the Mamdani fuzzy logic method were able to determine the conditions for push notifications to appear using the parameters as desired; by so doing, it is possible to create a battery monitoring system with accurate push notification feature to prevent the battery from being overcharged and overheated
A novel recommender system for adapting single machine problems to distributed systems within MapReduce
Full text linkThis research introduces a novel recommender system for adapting single-machine problems to distributed systems within the MapReduce (MR) framework, integrating knowledge and text-based approaches. Categorizing common problems by five MR categories, the study develops and tests a tutorial with promising results. Expanding the dataset, machine learning models recommend solutions for distributed systems. Results demonstrate the logistic regression model's effectiveness, with a hybrid approach showing adaptability. The study contributes to advancing the adaptation of single-machine problems to distributed systems MR, presenting a novel framework for tailored recommendations, thereby enhancing scalability and efficiency in data processing workflows. Additionally, it fosters innovation in distributed computing paradigms