Indonesian Journal of Electrical Engineering and Computer Science
Not a member yet
9109 research outputs found
Sort by
Analysis of similarity index between iThenticate and Ouriginal plagiarism detection software: a comprehensive study
Get PDFIntellectual property plagiarism is increasingly prominent in contemporary society, involving the unethical practice of claiming someone else's ideas, words, or creative works without proper acknowledgment. This study aimed to compare the performance of iThenticate and Ouriginal plagiarism detection software by analyzing their similarity index. Twenty original manuscripts (N=20) were examined for content similarity, with each manuscript analyzed first with Ouriginal and then with iThenticate. The focus was on comparing the two tools based on matched sources, word matches, and overall similarity index percentage. Data analysis using SPSS v26 included descriptive statistics, an independent t-test, correlation, and ranking of the similarity percentages, with significance set at p<0.05. The results indicated no significant differences in matching sources, matching words, or similarity index (p>0.05) between iThenticate and Ouriginal. A strong positive correlation (r=.758, p<.000) was observed between the similarity indices of the two software programs. The analysis of the low similarity range (≤10%) also revealed no statistical significant difference (p>.05). However, the mean similarity percentage detected by iThenticate was higher at 11.40%, compared to 6.85% for Ouriginal. Based on the findings, both iThenticate and Ouriginal demonstrated comparable effectiveness in detecting plagiarism, highlighting their importance in curbing academic dishonesty and protecting intellectual property rights
OFDM/CDMA channel quality estimation using K-means algorithm
Get PDFThis work aims to estimate the transmission channel quality and suggest a possible way to enhance the data rates to satisfy the increasing demand for higher data rates to a certain extent. The combination of any non-orthogonal subcarrier multiplexed (SCM) with CDMA needs a large bandwidth, hence a limited number of subcarriers and number of users as well as lower data rates. In contrast, orthogonal subcarriers such as the case of OFDM which are closely spaced due to their orthogonality property as well as to their reduced frequency selectivity fading are, therefore, crucial for increasing subcarriers and thus, increasing the data rates as well as the number of users. To describe the OFDM/CDMA technique in more detail, we performed a simulation using the software Scilab 5.5.2. In this simulation, we treat a simple example of a certain number of users using a bipolar orthogonal code, particularly, the Hadamard/Welsh code for the OCDMA, and the fast fourier transform (FFT) algorithm for the OFDM. For a more realistic simulation, we have introduced a gaussian white noise in the transmission channel and studied the effect of this noise on the eye diagram. Finally, to avoid the computational complexity in calculating the BER to study the OFDM/CDMA channel system quality, we have instead computed the bias and the variance of a noisy 16- quadrature amplitude modulation (QAM) constellation at the reception using the K-means algorithm
Enhancing sales volume using machine learning algorithms
Get PDFIn today's highly competitive business landscape, companies face a significant challenge in making accurate decisions based on vast amounts of historical data. Reliance on human data analysis often leads to biases and errors, hindering the ability to extract effective insights for sales forecasting. To address this challenge, this research presents an advanced model that integrates 14 machine learning (ML) regression algorithms, including XGBRegressor and LGBMRegressor, to provide accurate sales predictions using a comprehensive global store dataset. The results demonstrate that XGBRegressor and LGBMRegressor achieved the highest test accuracy (92%) and the lowest error rates, proving their ability to handle complex prediction tasks efficiently. This high accuracy in sales forecasting enables companies to make more effective strategic decisions, such as optimizing inventory management, allocating resources optimally, and exploring new growth opportunities. Consequently, the use of these advanced algorithms directly contributes to increasing sales volume and achieving a sustainable competitive advantage
A systematic evaluation of pre-trained encoder architectures for multimodal brain tumor segmentation using U-Net-based architectures
Get PDFAccurate brain tumor segmentation from medical imaging is critical for early diagnosis and effective treatment planning. Deep learning methods, particularly U-Net-based architectures, have demonstrated strong performance in this domain. However, prior studies have primarily focused on limited encoder backbones, overlooking the potential advantages of alternative pretrained models. This study presents a systematic evaluation of twelve pretrained convolutional neural networks—ResNet34, ResNet50, ResNet101, VGG16, VGG19, DenseNet121, InceptionResNetV2, InceptionV3, MobileNetV2, EfficientNetB1, SE-ResNet34, and SE-ResNet18—used as encoder backbones in the U-Net framework for identification and extraction of tumor-affected brain areas using the BraTS 2019 multimodal MRI dataset. Model performance was assessed through cross-validation, incorporating fault detection to enhance reliability. The MobileNetV2-based U-Net configuration outperformed all other architectures, achieving 99% cross-validation accuracy and 99.3% test accuracy. Additionally, it achieved a Jaccard coefficient of 83.45%, and Dice coefficients of 90.3% (Whole Tumor), 86.07% (Tumor Core), and 81.93% (Enhancing Tumor), with a low-test loss of 0.0282. These results demonstrate that MobileNetV2 is a highly effective encoder backbone for U-Net in extracting tasks for tumor-affected brain regions using multimodal medical imaging data
Robot vision and virtual reality integration to help paralyzed patients mobility
Get PDFThis study aims to develop a device that can assist the mobility of paralyzed patients, enabling them to communicate with family and caregivers by integrating robot vision and virtual reality (VR). The method used to connect audio and visual data communication between robot vision and VR is by utilizing the robot operating system (ROS2) middleware communication node through topics over a wireless network. In this research, paralyzed individuals can maneuver based on the movement direction of robot vision, which is remotely controlled via a joystick through Bluetooth communication. The input devices used in this system include a camera, microphone, joystick, and ultrasonic sensors. The processing part uses a Raspberry Pi as the data processing center, and the output includes a DC motor, servo motor, speaker, 5-inch monitor, and headset. The results indicate that the integration of robot vision and VR can assist paralyzed individuals in communicating with family or caregivers at distances of up to 10 meters. This is due to the maximum joystick control range for moving the robot via Bluetooth communication being 10 meters. Furthermore, this study shows that the use of robot vision and VR can improve paralyzed patients’ motivation, supporting the medical field in patient care
Effect of binaural beat brainwave entrainment on brainwave ratios in students with learning difficulties
Get PDFThis study examined the impact of binaural beat brainwave entrainment (BB BWE) on cognitive function and learning performance (LP) in children aged 8-13 with learning difficulties. A group of 52 participants was divided into a test group (TG) receiving BB BWE for four weeks and a control group (CG) without intervention. Results showed significant improvements in the TG, with LP increasing by up to 78% by week 4 according to cognitive assessment methods. EEG data corroborated these findings, showing a 74% improvement in TG students’ performance. Favorable changes in Electroencephalography (EEG) ratios were observed, including decreased theta/beta and theta/alpha ratios and an increased alpha/beta ratio. Topographical EEG maps revealed more balanced brain activity patterns post-BWE. The CG showed no significant changes. Notably, performance in the TG declined after discontinuing BWE, suggesting the need for ongoing intervention to maintain benefits. These findings indicate that BB BWE could be an effective non-invasive method for enhancing cognitive function and learning capacity in individuals with learning difficulties. However, further research is needed to establish long-term effects and optimal application protocols
Optimizing resume information extraction through TSHD segmentation and advanced deep learning techniques
Get PDFThis research focuses on a significant factor in the natural language processing area, which is extracting information from unstructured textual data through efficient methods in order to pull useful insights and structured representations from this data. This research attempts to boost the effectiveness of information retrieval systems through computational analysis. This paradigm is explored in this work using question answering models in an extractive style, a modern information extraction approach, creating a new methodology combining the topic segmentation based on headings detection (TSHD) segmentation algorithm and deep learning methods. The TSHD algorithm breaks documents into sections in which certain topics are addressed. Refined extraction models are then used to process these disjoint segments leading to more accurate and contextjudicious extraction compared to naive whole-document extraction approaches. We empirically validate this approach using the stanford question answering dataset (SQuAD) 1.1 dataset, with a specific adaptation to resumes. Experimental results show that the performance metrics increase by 7.4% in exact match (EM) and by 7.8% in F1-score. This can be concluded from these results illustrating the feasibility of the proposed approach in the automated information extraction frameworks such as resume processing
IoT-based intelligent crop rotation and recommendation system
Get PDFTraditional farming practices often rely on manual monitoring and crop selection, leading to inefficient use of resources and limited crop diversification. This study addresses these issues through the development of an IoT-based intelligent crop rotation and recommendation system that automates crop monitoring, irrigation, and crop selection processes. The system integrates DHT11 and NPK sensors to measure temperature, humidity, soil moisture, and nutrient levels (N, P, K), with real-time data displayed on a web-based application interface. An automated irrigation and fertilizer subsystem with SMS notifications enhances user control and remote accessibility. A crop recommendation module using the Euclidean Distance algorithm analyzes soil-nutrient data to identify the most suitable crops for the next planting cycle. System evaluation based on the ISO/IEC 25010 software quality model indicated high functionality, usability, reliability, portability, and maintainability, with an overall weighted mean of 3.958 (Agree) and a cronbach’s alpha of 0.9585, signifying excellent reliability. The system demonstrates the potential of internet of things (IoT)- based technologies in promoting crop diversification, optimizing farm productivity, and advancing sustainable agricultural practices
Panic detection through facial recognition paradigm using deep learning tools
Get PDFRecently, panic detection has become essential in security, healthcare, and human-computer interaction. Automatic panic detection (APD) systems are designed to monitor physiological signals and behavioral patterns in real-time to detect stress responses. APD is increasingly adopted across many sectors, including disaster preparedness, COVID-19, and terror attacks. Their integration with various applications reduces human efforts and saves costs. However, most studies rely on existing models with fewer new ones or techniques. This study proposes a vision-based panic detection model using MobileNet, ResNet, and convolutional neural network (CNN). The FER2013 dataset is used for the model training and testing. The results indicate that MobileNet is the most effective model for image-based panic detection across ten folds with an accuracy of 90%, recall of 96.9%, and mean accuracy of 0.032. MobileNet also showed a mean absolute error (MAE) between 0.02 and 0.04. This study has been to confirm MobileNet's suitability for image-based panic detection. The findings contribute to developing more reliable and accurate image-based panic detection systems in real-world applications. It offers valuable insights and lays the groundwork for future deep-leaning-based panic detection studies
Fuzzy multi-objective energy optimization of workflow scheduling
Get PDFTask scheduling is a key and challenging problem in cloud computing systems, requiring decisions regarding resource allocation to tasks to optimize a perfor mance criterion. This problem has required researchers and developers to over come significant challenges. Our goal in this study aims to minimize both the makespan and energy consumption in cloud computing systems by efficiently scheduling workflows. To achieve this, we first proposed a dynamic multi objective model, which wasthensimplified into a single-objective problem using dynamic weights. Then, we proposed a dynamic genetic algorithm (DGA) and a dynamic particle swarm optimization algorithm (DPSO) to address the prob lem. To deal with the situation where the makespan is uncertain and not exact, we present a fuzzy model, treating each value as a fuzzy number and we utilize both possibility and necessity metrics. The results are contrasted with the Het erogeneous earliest finish time (HEFT) algorithm and Considerably lowered the total energy consumption, especially for DGA