Bulletin of Electrical Engineering and Informatics
Not a member yet
2885 research outputs found
Sort by
Data-driven clustering of smart farming to optimize agricultural practices through machine learning
Get PDFThis study investigates the optimization of durian farming practices in Eastern Thailand using data-driven clustering techniques. The research aims to identify distinct agricultural patterns and improve resource allocation in durian production. K-means clustering is applied to durian production area and yield data from 2012 to 2023. Cluster quality is assessed using the Davies-Bouldin index (DBI), Dunn index, and Silhouette score. The methodology included comparing clustering results before and after log transformation of the data. Three main clusters are identified which are large-scale high-yield producers, small-scale lower-yield areas, and medium-scale producers with moderate yields. Notably, log transformation did not consistently improve clustering performance with original data often producing better-defined clusters. This finding highlights the importance of carefully considering data pre processing methods. Furthermore, the data-driven clustering offers valuable insights for precision agriculture by identifying regions with higher productivity allowing for targeted interventions and better resource allocation. The results can guide farmers in optimizing durian cultivation strategies, potentially leading to increased yields and more sustainable farming practices in Eastern Thailand's durian industry
Hybrid approach for tweets similarity classification founded on case based reasoning and machine learning techniques
Get PDFTwitter sentiment analysis becomes a popular research subject in the last decade. It aims to extract sentiments of users through their public opinion about a given topic. This article proposes a hybrid approach for Twitter sentiment analysis founded on dynamic case based reasoning (DCBR), multinomial logistic regression machine learning algorithm and multi-agent system. Our approach proposes a method to find similar tweets based on content similarity measure using the scientific measurement of keyword weight term frequency-inverse document frequency (TF-IDF). This approach includes gathering and pre-processing tweets, getting score and polarity of tweets, the use of multinomial logistic regression machine learning algorithm to classify our tweets into various classes, using the feature extraction method to extract useful features and then the K-nearest neighbors (KNN) algorithm to make it easier to find similar tweets to our tweet target case. This approach is adaptive and generic and able to track users' tweet to predict their behavior and sentiments in critical situations and delivering personalized content. The current study focuses on Covid-19 tweets, and a public Twitter dataset is used for this purpose
Optimizing earthquake damage prediction using particle swarm optimization-based feature selection
Get PDFEarthquakes have destroyed the economy and killed many people in many countries. Emergency response actions immediately after an earthquake significantly reduce economic losses and save lives, so accurate earthquake damage predictions are needed. This research looks at how machine learning (ML) techniques are used to predict damage from earthquakes. The ML algorithms used are k-nearest neighbors (KNN), decision tree (DT), random forest (RF), and Naïve Bayes (NB). Feature selection is necessary, it needs to select the most relevant features from big data. One of the most commonly used algorithms to optimize ML is particle swarm optimization (PSO). PSO is also suitable for feature selection. This research compares various of PSO. Based on research, the RF algorithm with Phasor PSO has the highest fitness score. This process succeeded in reducing features from 38 features to 14 features. Based on the process after feature selection, it was found that the KNN, DT, and RF algorithms had improved. RF obtained the best accuracy, namely 72.989%. The processing time in DT, RF, and NB is faster than before. In conclusion, the ML algorithm can be combined with PSO feature selection to create a classification model that provides better performance than without feature selection
Deep hybrid neural network for automatic classification of heart arrhythmias using 12-lead electrocardiograms
Get PDFThis research introduces a novel convolutional neural network-bidirectional long short-term memory (CNN-BiLSTM) hybrid network for the automatic classification of heart arrhythmias using 12-lead electrocardiograms (ECGs). By merging the spatial feature extraction capabilities of CNNs with the temporal precision of BiLSTM networks, our approach sets a new standard in cardiac diagnostics. The proposed model was tested against the comprehensive CPSC2018 dataset, demonstrating superior performance with an accuracy of 90.67%, precision of 93.27%, recall of 96.35%, and an F-score of 94.78%, surpassing existing state-of-the-art methods. These results underscore the effectiveness of integrating spatial and temporal data analysis, offering a robust and reliable tool for medical practitioners. This study represents a significant advancement in automated ECG analysis, paving the way for improved diagnosis and treatment of heart diseases, and contributing to enhanced patient outcomes in cardiac care
Integration of genetic algorithm and mesoscopic modeling for the optimization of membrane separation processes
Get PDFThis article is dedicated to the development of an innovative approach to optimizing membrane separation processes. The paper introduces the integration of a genetic algorithm (GA) and mesoscopic modeling to enhance the efficiency and accuracy of process parameter optimization. The GA is employed for evolutionary search of optimal parameters, such as pressure, temperature, and membrane material characteristics. The use of evolutionary principles allows for efficient exploration of parameter space, identifying optimal solutions. Mesoscopic modeling serves as a tool for detailed analysis and visualization of membrane separation processes. It involves modeling the interaction of molecules with the membrane surface, enabling a more accurate consideration of the physicochemical aspects of the process. The integration of the GA and mesoscopic modeling creates a unique tool for membrane separation process optimization. The developed approach contributes not only to improving component separation efficiency but also to minimizing energy consumption. The method presented in the article has been successfully tested on model membrane process systems and demonstrated significant improvements compared to traditional optimization methods. The research results confirm the potential of the proposed approach for application in membrane technology industries, opening new perspectives in the field of separation process optimization
Arabic dialect classification using an adaptive deep learning model
Get PDFIn daily life, dialect is the most widely used form of communication. Automatically identifying a dialect is a challenging task, particularly when dealing with similar dialects spoken in the same nation. In this study, we developed an automatic dialect identification of feature extraction based on the deep learning model. First, we extract the cepstral features, the fundamental frequency and glottal instances using our multi-scale product analysis (MPA) of the speech signal. These parameter measurements from the MPA of the speech signal are used as features for the designed Hamilton neural network (HNN) classifier. Our classifier considers both the external and the internal dependencies and allows one to code the dependencies by composing the multi-dimensional features as single entities as well as by determining the correlations between the elements by the recurrent operation. Experimental results show that the proposed dialect identification system achieves significant performance gains compared to current HNN-based approaches. The proposed system is rigorously designed to exploit the strong temporal and spectral relationships of speech, and its components operate independently and in parallel to accelerate processing. In addition, the experimental results indicated the robustness of our deep learning model for the identification of Arabic dialect
POA-DT: a novel method for predicting air quality in major Indian cities
Get PDFAir pollution is a critical environmental and public health concern, exacerbated by urbanization, industrial growth, and increased transportation. The air quality index (AQI) in major cities is significantly elevated due to rapid industrial expansion, fossil fuel consumption, and vehicular emissions. This study aims to predict AQIs using machine learning techniques, specifically integrating the Pelican optimization algorithm (POA) with the decision tree (DT) method to enhance accuracy. Data from prominent Indian cities—Mumbai, Delhi, Bangalore, Kolkata, and Chennai—was analyzed due to their high pollution levels. The model’s performance was validated against traditional machine learning methods such as k-nearest neighbors (KNN), random forest (RF) regression, and support vector regression (SVR). Results showed the highest prediction accuracies for Kolkata at 96.68%, followed by Bangalore at 95.66%, Chennai at 93.10%, Mumbai at 92.48%, and Delhi at 86.61%. These findings demonstrate that the proposed model outperforms conventional techniques in predicting AQI, providing a foundation for effective policy-making to mitigate air pollution impacts
The implementation of the K-nearest neighbor algorithm to detect the KRSRI robot obstacles
Get PDFThe Indonesian SAR robot contest (KRSRI) is a development of the fire extinguisher robot contest (KRPAI); initially, the robot at KRPAI only put out fires. Still, at KRSRI, the robot was asked to prioritize the SAR function. The robot had to overcome obstacles in this contest to complete it. Based on this, an obstacle detection system for the robot was designed using machine learning with the K-nearest neighbor algorithm and gray level co-occurrence matrix feature extraction. Later, the robot is expected to be able to carry out accurate obstacle detection to prioritize efficiency so that no more time is consumed due to the robot incorrectly detecting an obstacle. The results of the tests that have been carried out show that the detection accuracy based on the test dataset is 80% for rising barriers, 100% for debris obstacles, and 90% for step obstacles, and an error value of 20% for increasing obstacles is obtained, 0% for debris obstacles, and 10% for stair obstacles
Digital signal processing for accurate body temperature measurement
Get PDFThe relevance of the subject matter is conditioned by the need to find the most effective and stable point for measuring body temperature using special measuring instruments. The main purpose of this scientific research is to develop a method for the most accurate implementation of these operations through the use of scientific and theoretical apparatus of digital signal processing. The study used methods of analysis and synthesis of information about the main provisions necessary in the process of finding an effective place for measuring the temperature parameter. The main points of body temperature measurement were established, including the key provisions of the digital temperature signal processing technique for measuring these parameters, which are required to assess the real state of a person. The presented mathematical formulas reflect the possibilities of calculating the average body temperature, as well as the skin temperature parameter, determined depending on the temperature indicators at individual points. Temperature curves for various types of diseases were presented. The practical significance of the results obtained lies in the possibility of their use in the activities of institutions of various profiles to find an effective method of measuring human body temperature to assess the physical condition
Design and development of harmonic filters for harmonics reduction in polluted distribution network
Get PDFRecently due to development in the power electronics sector, there is a tremendous increase in nonlinear loads. These nonlinear loads cause distortion in the system current and result in degrading quality of power. The poor power quality causes technical and financial losses in the system which necessitates adoption of techniques to reduce the harmonic distortion to meet IEEE-standards and improve system efficiency. As per literature, passive, active and hybrid filter techniques are implemented to mitigate the harmonics. Each has merits and demerits. Constructive reduction in current harmonics improves the life and efficiency of equipment’s also assists to improve power quality and relieves penalties imposed by utilities. In this work, an attempt is made to give a detailed approach used in the designing of harmonic filters. This study will provide a broad outline to the engineer, researcher and consultant working in the field of power quality to design filters for the case under study. The steps to design the filters are well explained with mathematical equations and examples for greater insight. To validate the performance of the filter a MATLAB/Simulink platform is utilized. The outcome of the simulation result proved that current harmonics are minimized with a substantial amount