Bulletin of Electrical Engineering and Informatics
Not a member yet
2885 research outputs found
Sort by
Integrating multi-criteria decision making and reinforcement learning for consensus protocol selection
Get PDFThe rapid progress in artificial intelligence technologies in recent years has been largely driven by advances in reinforcement learning (RL). RL methods have proven to be highly effective in solving many practical problems. Distributed ledger technologies are finding wide application in the internet of things (IoTs), providing new approaches to solving problems of traditional IoT systems. Consensus is a fundamental component of distributed ledger technologies, responsible for ensuring data consistency between nodes, its security and accuracy. This paper is devoted to the study of the optimal choice of blockchain consensus protocol for IoT networks based on a combination of multi-criteria decision making (MCDM) and RL methods. The paper discusses the potential of merging MCDM and RL methods for selecting blockchain consensus protocols in IoT networks. It suggests a combined framework for effective protocol selection and management
Application of JAYA algorithm for optimizing allocation and size of thyristor-controlled series compensator devices
Get PDFElectricity serves as the backbone and essential energy source for various sectors, including transportation, residential areas, manufacturing, and industry. As engineering and technology advance, the demand for electricity continues to rise. Expanding the electricity grid to meet transmission needs and provide high-quality service has become a fundamental challenge in the power system domain. However, load expansion introduces issues such as line overloads when demand surges, compromising power quality, system security, and reliability during operation, potentially leading to system failures. Addressing these load-related problems is crucial for enhancing power system stability, reducing troubleshooting expenses, and improving operational efficiency. This study proposes the utilization of thyristor-controlled series compensator (TCSC) as a solution to enhance power system efficiency. Furthermore, to optimize TCSC placement and determine the appropriate compensation level for devices on transmission lines, the research suggests employing the JAYA optimization algorithm. MATLAB software is utilized to investigate the IEEE standard 30-node transmission lines case. The obtained results have demonstrated the effectiveness of the solution in enhancing electrical transmission capacity, improving stability, and reducing energy losses within the system at a low operational cost
Soybean leaf disease detection and classification using deep learning approach
Get PDFIn Ethiopia, where soybeans are mainly involved, manual observation has traditionally been relied upon for detecting soybean leaf diseases. However, the manual process is susceptible to numerous issues such as labor-intensiveness, inconsistency, and subjectivity. While previous studies have explored automated classification for soybean leaf disease detection, they primarily focused on binary classification, overlooking the complexity and diversity of soybean leaf diseases, which hinders effective management strategies. This study introduces deep learning algorithms and computer vision for automated soybean leaf disease identification and classification in soybean leaves. By comparing pre-trained convolutional neural network (CNN) models (VGG16, VGG19, and ResNet50V2), a dataset of 3078 soybean leaf images was curated, representing various diseases. Image preprocessing techniques augmented the dataset to 6,958 images, enhancing the model's accuracy and generalization performance. VGG16 demonstrated outstanding performance with a test accuracy of 99.35%, highlighting its promising performance and generalization potential
Solving missing categorical data in questionnaire responses for automated classification
Get PDFHandling missing categorical data is critical for maintaining the accuracy and reliability of automatic classification tasks, particularly in mental health screening based on questionnaire responses. This study investigates several imputation methods, including last observation carried forward (LOCF), k-nearest neighbor (KNN) imputation, hot-deck imputation, and multivariate imputation by chained equations (MICE). Results show that KNN imputation achieves the lowest root mean square error (RMSE), indicating the most faithful reconstruction of the original data. However, for classification performance, MICE-imputed datasets produced models that outperformed those generated by other methods and even surpassed models trained on the original incomplete data. Interestingly, we also found that using observed data over multiple iterations of imputation tuning can introduce greater deviation from original missing values, but this process can help form datasets with clearer class boundaries, ultimately improving classification accuracy. These findings emphasize the need to balance data fidelity and model performance when selecting imputation strategies
Benchmarking machine learning algorithm for stunting risk prediction in Indonesia
Get PDFStunting is a condition caused by poor nutrition that results in below-average height development, potentially leading to long-term effects such as intellectual disability, low learning abilities, and an increased risk of developing chronic diseases. One effort to reduce stunting is to apply a machine learning algorithm with a data science approach to develop risk prediction models based on factors in stunting. The study used the current cross industry standard process for data mining (CRISP-DM) framework to gain insight and analyzed 1561 records of data collected from the Indonesia family life survey (IFLS) for the prediction models. Two sampling methods, random undersampling, and oversampling synthetic minority oversampling technique (SMOTE), were employed and compared to overcome the data imbalance problem. Four machine learning classifier algorithms were trained and tested to determine the best-performing model. The experiment results showed that the algorithms yielded an average accuracy of more than 75%. Using the undersampling technique, the accuracy obtained by logistic regression, k-nearest neighbor (KNN), support vector classifier (SVC), and decision tree classifier were 95.21%, 78.91%, 92.97%, and 86.26% respectively. Meanwhile, the oversampling technique reached 96.17%, 88.50%, 93.29%, and 95.21%, respectively. Logistic regression emerges as the best classification, with oversampling yielding superior performance
Development of a robust and sustainable regional demography-based demand management technique
Get PDFThis paper presents a robust and sustainable energy management system driven by regional demographic patterns developed using fuzzy logic and mixed integer linear programming (MILP). This method detects and integrates variations in the energy use patterns of urban and rural communities attaining improved efficiency in the management of regional power demand. The detection and integration of the urban and rural energy use patterns were done by combining period partitioning based regional time of use tariff and fuzzy based appliance level renewable resource allocation to develop a function to be optimized using an improved MILP which provides users with the optimum schedule of appliance usage based on their demographic classification. The effectiveness of the proposed method was tested by running MATLAB simulations of different scenarios emulating continuous regional renewable integration planning with urban and rural power consumption profiles generated using LoadProGen. The proposed method’s effectiveness is confirmed by the achievement of a reduction upto 31% in the community energy cost as well as significant reduction in the energy costs of each participant over different scenarios compared to the unoptimized base case. The proposed method can be effectively utilized in energy management applications catering to multiregional and mixed demographic communities
Performance comparison of algorithms in the classification of fresh fruit types based on MQ array sensor data
Get PDFAccurate classification of fresh fruit types is essential in the agricultural sector for ensuring quality control, minimizing waste, and enhancing food safety across the supply chain. This study evaluates the performance of four machine learning algorithms—artificial neural network (ANN), K-nearest neighbors (KNN), logistic regression (LR), and random forest (RF)—in classifying fruit freshness based on data obtained from electronic noses equipped with MQ array sensors. Experiments were conducted using a comprehensive dataset comprising various fruit combinations, and model performance was assessed using accuracy, precision, recall, and F1 score metrics. Results indicate that the RF algorithm achieved the highest accuracy (100%) and precision (1.00), demonstrating superior performance in both classification accuracy and computational efficiency. ANN and KNN also performed well, with accuracies of 96.80% and 97.10%, respectively, while LR yielded a lower but still effective accuracy of 91.16%. Statistical analysis confirms that RF's superior performance is statistically significant when compared to the other algorithms. These findings suggest that RF is the most effective algorithm for fruit freshness classification using electronic nose data, offering fast and reliable results that are well-suited for integration into real-time monitoring systems in agricultural and food retail applications
Incremental learning based fuzzy reasoning approach for diagnosis of thyroid disease
Get PDFThis paper presents a novel hybrid fuzzy logic approach for the classification of thyroid disease. Hybrid fuzzy logic approaches have brought many benefits to the medical data classification problems such as reasoning on uncertain or incomplete data. The machine learning algorithms had been used with the fuzzy expert systems to define the fuzzy rule base. The optimization techniques had been used in the fuzzy expert systems for optimizing the fuzzy membership functions and fuzzy rules. Enhancing the machine learning algorithms and optimization techniques that are integrated with the fuzzy logic method can improve the overall performance of the fuzzy expert system. To deal with the curse of dimensionality problem and to enhance the integration of machine learning algorithm and fuzzy logic method, this paper presents an incremental learning based parallel fuzzy reasoning system (IL-PFRS) for medical diagnosis. In this research work, the decision tree classifier is used to extract the features from dataset. IL-PFRS is applied to classify the thyroid disease which is serious disease that needs attention and earlier detection. The thyroid disease dataset obtained from the UCI machine learning repository is used in this research work where the IL-PFRS showed the classification accuracy of 99% when testing using this dataset
Enhanced real-time glaucoma diagnosis: dual deep learning approach
Get PDFEffective management of glaucoma is essential for preventing irreversible vision loss. This study introduces a novel deep learning-based network designed to enhance performance while minimizing computational complexity. The system comprises two models: the first is a hybrid model combining a customized U-Net architecture integrated with you only look at coefficients (YOLACT) is utilized to achieve accurate segmentation of the optic disc (OD) and optic cup (OC), providing detailed diagnostic insights for ophthalmologists. The second model employs you only look once version 5 (YOLOv5) for real-time glaucoma prediction, delivering outstanding performance with an accuracy of 97.89% and F1 score of 98% on the primary dataset. On an independent dataset without further training, the model achieved 96% accuracy, with sensitivity and specificity of 98.9% and 93.3%, respectively. These results highlight the model's robustness, generalizability, and adaptability, demonstrating its potential for effective glaucoma screening and early detection in diverse clinical environments. This approach offers a promising advancement in improving the accessibility and efficiency of glaucoma management
Internet of things forensic: contemporary issues, challenges, and future research directions
Get PDFThe internet of things (IoT) comes with great capabilities as well as new opportunities for attackers and criminals. Even though digital forensics is considered to be mature and has been studied by many researchers in recent years, IoT forensics is relatively new and yet to be thoroughly explored. IoT technology has unique characteristics that require adoption of the traditional digital forensic approaches, frameworks, and tools. The primary goal of IoT forensics is to collect, preserve, and present evidence in a manner that meets legal standards and country law from a specific IoT system that includes connected devices and sensors via different types of networks and associated cloud environments. In this paper, we explored the main difference between IoT forensics and traditional digital forensics. This paper aims to provide a comprehensive and up-to-date overview of recently proposed solutions for addressing the IoT forensic domain. we highlighted the limitations of the recently proposed framework and the utilized technologies by researchers. In addition, we recommended some new research directions that could enhance the IoT investigation process. The goal of this paper is to provide a clear understanding of the currently used technologies and other fields in IoT forensic frameworks, limitations, and directions in this area, which may be helpful for future researchers interested in this field