Indonesian Journal of Electrical Engineering and Computer Science
Not a member yet
9109 research outputs found
Sort by
A novel mobile application for personality assessment based on the five-factor model and graphology
Get PDFWith the rising interest over the last decade, automated graphology has emerged as a promising filed of research, providing new insights on personality traits prediction on the basis of handwriting analysis. Although, few practical solutions to automate the extraction of handwriting features and personality prediction exist in the literature. This work aims to contribute to closing the gap in automated handwriting personality prediction by proposing a novel mobile application that uses robust feature extraction and machine learning models to predict big five personality traits. Our findings, based on high correlations between handwriting characteristics and personality traits, revealed convincing links. Notably, extraversion and extraversion have strong correlations with top margin feature, whereas agreeableness is expressed through line spacing. These findings emphasize the ability of automated graphology to properly interpret individual personalities. The proposed system achieved exceptional accuracy by using well known machine learning classifiers. The testing accuracy exceeded 92% in binary classification and 87% in multi-class case scenario, proving the adaptability and dependability of the system’s architecture. Our Android app promises to provide users with unprecedented insights into their personalities, establishing a robust tool for psychological assessment and self-discovery
Word embedding for contextual similarity using cosine similarity
Get PDFPerspectives on technology often have similarities in certain contexts, such as information systems and informatics engineering. The source of opinion data comes from the Quora application, with a retrieval limit of the last 5 years. This research aims to implement Indo-bidirectional encoder representations from transformers (BERT), a variant of the BERT model optimized for Indonesian language, in the context of information system (IS) and information technology (IT) topic classification with 414 original data, which, after being augmented using the synonym replacement method, The generated data becomes 828. Data augmentation aims to evaluate the performance of models by using synonyms and rearranging text while maintaining meaning and structure. The approach used is to label the opinion text based on the cosine similarity calculation of the embedding token from the IndoBERT model. Then, the IndoBERT model is applied to classify the reviews. The experimental results show that the approach of using IndoBERT to classify SI and IT topics based on contextual similarity achieves 90% accuracy based on the confusion matrix. These positive results show the great potential of using transformer-based language models, such as IndoBERT, to support the analysis of comments and related topics in Indonesian
Internet of things based autonomous robot system architecture for home automation and healthcare services
Get PDFThe internet of things (IoT) is playing a major role in the development of the health industry by enabling more accessible and affordable virtual and distant patient contacts through applications that are easy to use. The IoT and automated homes are becoming more popular in recent days. A network of connected devices, including hardware, equipment, and technical support, is known as the IoT. Their purpose is to allow data exchange with other systems through the internet. This paper presents, internet of things based autonomous robot system architecture (IoT-ARSA) for home automation and healthcare services. The primary goal of this secure home automation system is to help the elderly and disabled people by allowing them to operate home appliances. Additionally, the system uses a cloud server to predict the health conditions of patients and the elderly people, providing information to a guardian. The patient's health condition is determined using sensors like temperature, pulse, blood pressure, and oxygen level. Ultrasonic sensor and face detection are used for home automation. Each sensor will interact with the Raspberry Pi 4 to record data, which will then be processed and stored in the cloud. From results it is clear that described (IoT-ARSA) for home automation and healthcare services model is very efficient with high accuracy and high security. Health monitoring is achieved with this model continuously with great efficiency
Chebyshev distance-embedded twin support vector machine for skewed classification problems
Get PDFSupport vector machine (SVM) is a pivotal classification algorithm, and its evolutionary counterpart, the twin SVM (TWSVM), has gained acclaim for its advanced generalization capabilities, particularly in handling imbalanced data. TWSVMs achieve swift training by explicitly exploring a pair of non-parallel hyperplanes, yet selecting numerical values for hyperparameters poses a challenge due to the uncertainty introduced by random preferences. This paper presents a novel approach, the Chebyshev distance-based TWSVM, specifically designed for hyperparameter tuning in imbalanced binary classification. This innovative model mitigates the uncertainty of hyperparameter selection by leveraging Chebyshev distance, thereby enhancing the generalization capabilities of the TWSVM. To evaluate its efficacy, computational tests were conducted on publicly accessible real-world benchmark datasets across various domains, including non-linear cases. The results demonstrate that the Chebyshev distance-based TWSVM outperforms several existing methods, achieving superior performance with reduced computational time and setting a new benchmark in the field
Characterization of binarized neural networks for efficient deployment on resource-limited edge devices
Get PDFThis paper delves into binarized neural networks (BNNs) tailored for resource-constrained edge devices. BNNs harness binary weights and activations to amplify efficiency while upholding accuracy. Across diverse network configurations, BNNs consistently outshine traditional neural networks (NNs). A pioneering BNN architecture is developed in LARQ, achieving an impressive. 61% accuracy on the MNIST dataset through binary quantization, weight clipping, and pointwise convolutions. Implementation on the Xilinx PYNQZ2 FPGA board shows far quicker classification rates, with a maximum inference time of 0.00841 milliseconds per image, approximately 10,000 images being classified in this length of time. The time taken per image represents approximately 0.01% of the total inference time. This underscores BNNs' potential to redefine real-time edge computing applications. The paper makes significant strides by elucidating BNNs' performance superiority, proposing an innovative architecture, and validating its prowess through real-world deployment. These findings underscore BNNs as agile, high-performance models primed for edge computing, fostering a new era of real-time processing innovations
Binary white shark optimization algorithm with Z-shaped transfer function for feature selection problems
Get PDFFeature selection is critical for improving model performance and managing high-dimensional data, yet existing methods often face limitations such as inefficiency and suboptimal results. This study addresses these challenges by introducing a novel approach using the white shark optimization (WSO) algorithm and its binary variants to enhance feature selection. The proposed methods are evaluated on various datasets, including “Dorothea,” “Breast Cancer,” and “Arrhythmia,” focusing on classification accuracy, the number of features selected, and fitness values. Results demonstrate that the WSO algorithms significantly outperform traditional methods, offering notable improvements in accuracy and efficiency. Specifically, the WSO variants consistently achieve higher accuracy and better fitness values while effectively reducing the number of selected features. This research contributes to the field by providing a more effective optimization approach for feature selection, addressing existing inefficiencies, and suggesting future directions for further refinement and broader application. The findings highlight the potential of advanced optimization techniques in enhancing data analysis and model performance, offering valuable insights for practitioners and researchers
Creating inclusive UX: uncovering gender-bugs in higher education website through GenderMag’ing
Get PDFHigher education websites serve as service-providing and information-disseminating platforms which may contain gender-related usability issues that affect how male and female users interact with digital platforms. This study applied the gender inclusiveness magnifier (GenderMag) method to identify and assess these gender-specific usability barriers. Researchers conducted cognitive walkthrough sessions using gendered personas, Abi (female) and Tim (male), uncovering key inclusivity bugs aligned to specific cognitive facets-motivation, information processing style, computer self-efficacy, risk aversion, and learning style. Insights from these walkthroughs guided the creation of a structured usability survey, administered to 200 respondents equally divided between males and females, comprising faculty and upper-year BS information technology students. Statistical analysis revealed significant gender differences specifically in information processing style (p=0.0003), emphasizing distinct preferences for content organization and navigation between genders. The integration of usability factors with GenderMag’s cognitive facets effectively pinpointed areas requiring inclusive design adjustments, guiding future efforts to enhance equitable digital interactions in educational environments
The road conditions detection using the convolutional neural network
Get PDFPoor road conditions present considerable obstacles for individuals, resulting in asset loss, bodily harm, and time inefficiency. Approximately 1.35 million fatalities are attributable to road traffic incidents. The Department of Public Works and Town & Country Planning conducted road surveys to assess and strategize maintenance efforts. The manual car survey requires additional time and an excessive budget. The automated system of artificial intelligence (AI) is widely recognized. This paper presents a model to detect road surface conditions utilizing video data. Four versions of convolutional neural networks (CNN) were utilized for this work. The model evaluation employed the mean average precision (mAP) measure. The video data was acquired via a smartphone mounted in a vehicle, comprising 10,984 photos for training and 2,198 images for testing. We trained and evaluated four versions of CNN architectures named YOLO, utilizing our data and GPU, with a specific emphasis on identifying cracks, potholes, and the condition of manhole covers. Of the architectures evaluated, YOLO V6 attained the greatest mAP score in comparison YOLO V5 to YOLO V8. The testing results with batch sizes of 4, 8, 16, and 32 are effective. The batch size of 32 yields the highest performance, achieving 87.38% mAP. Conduct the dropout normalization using rates of 0.25, 0.50, 0.75, and 1. The maximum mAP is observed with a dropout rate of 0.25, yielding a mAP of 85.40%. The model indicates that the government conducted road surface inspections with enhanced efficiency, enabling the planning of road repairs for public utility issues, which can lower transportation costs. Additionally, the model can be utilized to identify hazardous road conditions and minimize vehicular accident rates
Web-Based Attacks Detection Using Deep Learning Techniques: A Comprehensive Review
Get PDFWeb applications are utilized extensively by a broad user base, and the services provided by these applications assist enterprises in enhancing the quality of their service operations as well as increasing their revenue or resources. To gain control of web servers, attackers will frequently attempt to modify the web requests that are sent by users from web applications. Attacks that are based on the web can be detected to help avoid the manipulation of web applications. In addition, a variety of research has offered many methods, one of which is artificial intelligence (AI), which is the method that has been utilized the most frequently to identify web-based attacks recently. When it comes to the protection of web applications, anomaly detection techniques used by intrusion prevention systems are preferred. Deep learning, often known as DL, is going to be covered in this paper as anomaly-based web attack detection methods and machine learning techniques. With the purpose of organizing our selected techniques into a comprehensive framework that encourages future studies, we first explained the most concepts that related to web-based attacks detection, then we moved on to discuss the most prevalent web risks and may provide inherent difficulties for keeping web applications safe. We classify previous studies on detecting web attacks into two categories: deep learning and machine learning. Lastly, we go over the features of the previously utilized datasets in summary form
Enhancing marketing efficiency through data-driven customer segmentation with machine learning approaches
Get PDFThe importance of understanding consumer behavior in transaction data has become a key to improving marketing efficiency. This study aims to explore the application of machine learning (ML) techniques for data-driven consumer segmentation, focusing on improving product marketing strategies. This work addresses the limitations in the existing literature, especially in terms of handling high-dimensional data that can reduce segmentation quality. Previously, various studies have used clustering algorithms such as K-means without considering dimensionality reduction, which often leads to decreased accuracy and long computation time. In this study, we propose a new approach that combines principal component analysis (PCA) for dimensionality reduction and K-means clustering for consumer segmentation based on purchasing behavior. Experimental results show that using PCA to reduce data dimensionality significantly improves segmentation quality with an inertia score of 1,455,650 and a silhouette score of 0.486366. By implementing this method, we can group consumers into three segments based on frequently purchased product categories and the most common payment methods. These findings provide a scalable, data-driven segmentation framework that can be applied to improve marketing effectiveness by providing special discounts on various products based on the payment method used