cover
Contact Name
Christian Harito
Contact Email
christian.harito@binus.edu
Phone
+6221-5350660
Journal Mail Official
aagung@binus.edu
Editorial Address
Universitas Bina Nusantara Jl. Kebon Jeruk Raya No.27 Kebon Jeruk, Jakarta Barat 11530
Location
Kota adm. jakarta barat,
Dki jakarta
INDONESIA
Engineering, Mathematics and Computer Science Journal (EMACS)
ISSN : -     EISSN : 26862573     DOI : https://doi.org/10.21512/emacs
Engineering, MAthematics and Computer Science (EMACS) Journal invites academicians and professionals to write their ideas, concepts, new theories, or science development in the field of Information Systems, Architecture, Civil Engineering, Computer Engineering, Industrial Engineering, Food Technology, Computer Science, Mathematics, and Statistics through this scientific journal.
Articles 174 Documents
Adaptive Fuel Subsidy Optimization Using Deep Q-Learning and Bandit-Based Policy Selection: A Simulation Study Pambudi, Pandu Dwi Luhur
Engineering, MAthematics and Computer Science Journal (EMACS) Vol. 7 No. 2 (2025): EMACS
Publisher : Bina Nusantara University

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21512/emacsjournal.v7i2.13419

Abstract

Designing effective fuel subsidy policies is a major challenge for governments seeking to balance energy affordability, fiscal sustainability, and environmental goals. This study introduces an adaptive simulation framework combining Deep Q-Learning and a multi-armed bandit algorithm to model fuel consumption behavior and optimize subsidy distribution strategies. Moreover, this paper simulates a dual-agent system in which a DQN-based consumer interacts with a bandit driven government selecting among three subsidy policies: universal, quota-based, and targeted. By simulating consumer responses to universal, quota-based, and targeted subsidies over 1,000 episodes, the framework demonstrates how policy can adapt in real-time to maximize social welfare and reduce inefficient spending. Results show that while universal subsidies often deliver the highest consumer satisfaction, they incur significant fiscal costs, whereas quota and targeted approaches can yield more balanced trade-offs. The study highlights the potential of reinforcement learning to enhance adaptive policymaking in complex economic systems. To strengthen the analysis, the simulation tracks both consumer and government rewards across scenarios, capturing the trade-off between satisfaction and fiscal burden. Evaluation results reveal that targeted subsidies, though less popular, often provide more sustainable outcomes. The agent-based approach enables the system to dynamically adjust policy decisions based on real-time feedback, reflecting the evolving nature of economic behavior. These findings underscore the usefulness of AI-driven simulations as decision-support tools in designing responsive and cost-efficient public policies.
The Impact of Text Preprocessing in Sarcasm Detection on Indonesian Social Media Contents Jeremy, Nicholaus Hendrik
Engineering, MAthematics and Computer Science Journal (EMACS) Vol. 7 No. 2 (2025): EMACS
Publisher : Bina Nusantara University

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21512/emacsjournal.v7i2.13503

Abstract

Sarcasm is a way to convey something but delivered in the opposite way. This behavior is common on social media, where there are plenty of examples. On natural language processing, the task on its own is difficult primarily due to the lack of context. To add another layer of difficulty, communication in social media is done colloquially. One sacrasm benchmark, IdSarcasm, has alleviated one key issue in the development of sarcasm detection. However, there has not been an attempt to further preprocess the input before feeding them into the model. Pre-trained language models always use preprocessed corpus to ensure that the model is built upon quality dataset. Based on the current condition of IdSarcasm, further preprocessing step is necessary to ensure better quality. Specifically, the additional steps needed are handling HTML code, code-mixing, and colloquial writing which consists of shortened form, extended form, spelling variation, and reduplication. Several scenarios are created to observe the effect of additional preprocessing steps. Each additional preprocessing step is also tested to observe the effect of the preprocessing step independently. We prove that preprocessing step is still prevalent for data sourced from social media, and we recommend IndoNLU’s IndoBERT or large multilingual model to be used for sarcasm classification.
SMOTE Effectiveness and various Machine Learning Algorithms to Predict Self-Esteem Levels of Indonesian Student Anshori, Mochammad; Siwi Pradini, Risqy; Teja Kusuma, Wahyu
Engineering, MAthematics and Computer Science Journal (EMACS) Vol. 7 No. 2 (2025): EMACS
Publisher : Bina Nusantara University

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21512/emacsjournal.v7i2.13521

Abstract

Self-esteem plays a crucial role in students' psychological well-being, influencing their academic performance and personal development. Despite its importance, self-esteem is challenging to measure due to its abstract and subjective nature. This study aims to develop a predictive model to classify students’ self-esteem levels as high or low using machine learning and tabular data obtained through questionnaires. A dataset comprising 47 student responses, with 19 features consisting of social, emotional, demographic aspects, were analyzed. Five machine learning models were evaluated: Naïve Bayes, Decision Tree, Random Forest, Logistic Regression, and Support Vector Machine (SVM). To address the class imbalance in the dataset, the study applied SMOTE for data balancing and min-max normalization for feature standardization. Model performance was assessed using accuracy and F1-score. The results reveal that SVM, particularly with an RBF kernel, outperformed other models across all scenarios. On raw data, SVM achieved 66% accuracy and an F1-score of 57.3%. After applying SMOTE, the performance improved to 80% accuracy and a 79.9% F1-score. Further enhancement with normalization resulted in the best performance, with SVM achieving 83.33% accuracy and an F1-score of 83.3%. These results demonstrate how well preprocessing methods work to enhance machine learning models for datasets that are unbalanced. The proposed SVM-based model offers promising applications in educational and psychological settings, enabling early interventions to support students’ mental health.
Implementation of Microservices Architecture in a Retail Web Application Using Apache Kafka as a Message Broker Daeli, Stefanus; Lase, Kristian Juri Damai; Sumihar, Yoel Pieter
Engineering, MAthematics and Computer Science Journal (EMACS) Vol. 7 No. 2 (2025): EMACS
Publisher : Bina Nusantara University

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21512/emacsjournal.v7i2.13932

Abstract

Web-based applications are often initially developed using monolithic architecture due to its simplicity and ease of deployment. However, as application complexity grows, monolithic systems face critical limitations in scalability, flexibility, and performance. This research applies a microservices architecture to a Retail Web divided into four core services: user, product, transaction, and notification management. Apache Kafka is integrated as a message broker to support asynchronous, real-time communication across services. A total of 2,001 requests were recorded during system testing using Prometheus. The srv_tulityretailaccounts service achieved an average response time of 122.8 ms, and the srv_tulityretailtransactions service maintained 188.1 ms with a 98% success rate. The srv_tulityretailproducts service also demonstrated stable performance with consistently low response times and no error spikes. Meanwhile, the srv_tulityretailnotifications service showed the highest efficiency with an average response time of 28.5 ms, CPU usage at 12.75% (1.53 of 12 cores), and memory usage at 2.07 GB (56.5%) of 3.66 GB. Throughout testing, no service exhibited resource saturation or degradation, even under concurrent load conditions. This confirms the system’s horizontal scalability, where each service can independently scale without impacting others. Overall, the microservices approach has proven effective in enhancing performance, modularity, and production-readiness, while laying a strong foundation for continuous integration, deployment automation, and future feature expansion.
A Study of Machine Learning Approach to Predict the Out performance Market of Japan’s Stock Price Amadea, Charissa; Karina, Karina; Addina, Kanisha; Adara, Keisha; Pasaribu, Asysta
Engineering, MAthematics and Computer Science Journal (EMACS) Vol. 7 No. 2 (2025): EMACS
Publisher : Bina Nusantara University

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21512/emacsjournal.v7i2.12905

Abstract

This study employs machine learning algorithms to estimate the stock performance of Japanese companies in 2022, with a focus on examining the relationship between significant financial factors—Market Capitalization (Market Cap), Price-to-Earnings Ratio (PER), Price-to-Book Value (PBV), Return on Equity (ROE), and Debt-to-Equity Ratio (DER)—and stock performance, classified as either high-performing or low-performing. These factors are designated as independent variables. The dataset comprises 1,000 publicly listed companies in Japan and is analyzed using a logistic regression model. The dependent variable in this study is stock performance. The Akaike Information Criterion (AIC) guided model selection to optimize predictive accuracy and model complexity. The dataset was split into 70% for training and 30% for testing to ensure robust model validation. The best-performing model achieved a prediction accuracy of 62.67%, demonstrating strong sensitivity (88.83%) but weak specificity (18.75%). An AUC value of 0.6226 indicates moderate discriminatory power. The model shows good capability in detecting underperforming stocks, while its limitation lies in classifying well-performing stocks. The study suggests enhancing prediction accuracy by incorporating additional relevant variables such as macroeconomic indicators or market trends, as well as employing more complex machine learning algorithms like Random Forest or Gradient Boosting. These findings not only contribute to the literature on stock market prediction but also offer practical insights for investors in making investment decisions.
Sentiment Analysis of Slang Language Trends in Generation Alpha on Social Media Using BERT Priccilia, Shania; Erin, Erin
Engineering, MAthematics and Computer Science Journal (EMACS) Vol. 7 No. 2 (2025): EMACS
Publisher : Bina Nusantara University

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21512/emacsjournal.v7i2.13368

Abstract

Generation Alpha is a group growing up in an era of rapid digital technology advancement. Unlike previous generations who experienced a transition into technology, Generation Alpha demonstrates unique communication characteristics, particularly in their frequent use of slang, which is often difficult for older generations to understand. This gap in language understanding can lead to miscommunication, especially when the meaning of slang is misinterpreted. This phenomenon presents a challenge in establishing intergenerational communication, especially in digital and social media contexts where informal language is dominant. This study aims to explore the effectiveness of AI models in analyzing the sentiment of slang language used by Generation Alpha. Three BERT-based models were utilized in this research: BERT, RoBERTa, and DistilBERT. These models were selected based on their performance and efficiency in natural language processing (NLP) tasks, particularly in text classification and sentiment analysis. The dataset consists of 24,958 slang-based posts collected from users on the social media platform X. The analysis shows that DistilBERT achieved the highest accuracy score of 0.83, followed by BERT (0.82) and RoBERTa (0.81). These findings suggest that BERT-based models, especially DistilBERT, perform reliably in identifying the sentiment behind slang expressions used by Generation Alpha and hold potential for implementation in AI-based moderation or social media monitoring systems.
Wearable Sensors for Health Monitoring: Technologies, Applications, Challenges, and Future Perspectives Abrori, Syauqi Abdurrahman
Engineering, MAthematics and Computer Science Journal (EMACS) Vol. 7 No. 2 (2025): EMACS
Publisher : Bina Nusantara University

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21512/emacsjournal.v7i2.13765

Abstract

This article presents the state of the art and future outlook for wearable sensors for health monitoring, with emphasis on their roles in tracking physiological, biochemical, motion, and environmental parameters. Wearable sensors have moved beyond activity monitoring to facilitate clinical applications like chronic disease management, remote monitoring, and mental health evaluation. Four sensors are presented, with the sensing principle, formats, and actual-world application. System architectural elements like data acquisition, wireless communication, on-device and cloud processing, and user interface are addressed. The latest advancements like multi-modal sensor fusion, self-sustaining platforms, integration of machine learning, and skin-conformable electronics are also outlined. Wearable technology holds promise and is plagued with accuracy, battery life, privacy of data, and compatibility with health information systems. These hindrances need to be overcome if broader clinical integration and global accessibility are to take place. Avenues for development include energy-autonomous sensors, personalized feedback systems, and digital twin integration, which have promising potential for making early intervention, preventive care, and decentralized healthcare delivery possible. This overview provides a general background to researchers, developers, and clinicians striving for the next generation of digital health solutions.
Analysis of the Relationship Between Implementation and Policy for Regional-Scale Waste Management in Garut District Maulana, Suhenra
Engineering, MAthematics and Computer Science Journal (EMACS) Vol. 7 No. 3 (2025): EMACS
Publisher : Bina Nusantara University

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21512/emacsjournal.v7i3.12689

Abstract

Waste management is currently still a big challenge for all cities/regencies in Indonesia where there is still a lot of household waste that has not been managed and still relies on open dumping. This study aims to determine the relationship between the existence of a policy for the provision of regional-scale Waste Treatment Facilities with TPS 3R and its implementation in each village/sub-district in Garut District. The research method used is quantitative, and the data analysis used in this study is explanatory research which aims to explain whether or not there is a relationship between the independent variables, namely knowing the regent's instruction letter regarding the provision of regional-scale Waste Treatment Facilities with TPS 3R. The process of analyzing research data using the Fisher's Exact Test statistical test shows that the p-value is 0.158. Based on the applicable provisions, if the p-value is smaller than the significance level (α = 0.05) then H_0 (null hypothesis) is rejected and H_1 (alternative hypothesis) is accepted. Therefore, it was found that the p-value is greater than the significance level (0.158 > 0.05) then H_0 is accepted, which means that there is no statistically significant relationship between knowing the policy of providing regional-scale Waste Treatment Facilities with TPS 3R with the implementation of the provision of waste treatment facilities. This means that knowledge of the existence of policies through the regent's instruction letter does not have a significant effect on the implementation of waste treatment in villages/sub-district areas, this could mean that there are other factors that have a greater influence on the implementation of waste treatment in villages/sub-districts.
Proof of Data Weigher Analysis (DWA) and Its Application to Dynamic Meta Data Weigher Goenawan, Stephanus Ivan
Engineering, MAthematics and Computer Science Journal (EMACS) Vol. 7 No. 3 (2025): EMACS
Publisher : Bina Nusantara University

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21512/emacsjournal.v7i3.13155

Abstract

Data Weigher Analysis (DWA) addresses the persistent problem of objectively quantifying whether the values in a data set lean more heavily toward the left or right side, a challenge that becomes increasingly complex in irregular or large-scale data sets. Motivated by the need for a simple yet rigorous quantitative framework, this study compares two DWA techniques—the data weighting method and the data mean difference method—designed to compute balance points in a sequence. The data weighting method assigns balanced linear weights to left and right subsets, whereas the data mean difference method calculates first- and second-order mean differences to capture asymmetry in data distribution. We provide a theoretical proof of equivalence between these two formulations, showing that the mean difference approach produces the same linear weighting as the original data weighting scheme. Building on this theoretical result, we introduce a sliding-window algorithm to operationalize DWA on large, dynamic data streams, allowing automated detection of local imbalances in real time. Empirically, we validate our approach on real-world metadata and trade datasets, comparing it against baseline descriptive statistics to assess efficiency and precision. Quantitative findings show that the mean difference method reduces computation processes without loss of accuracy compared with manual weighting. Overall, this work contributes to a unified theoretical foundation, a lightweight algorithmic implementation, and evidence of practical benefits for using DWA in decision-making contexts such as questionnaire analysis, market dynamics, and trade flow monitoring.
Hybrid CNN-Based Classification of Coffee Bean Roasting Levels Using RGB and GLCM Features Halim, Rico; Riftiarrasyid, Mohammad Faisal
Engineering, MAthematics and Computer Science Journal (EMACS) Vol. 7 No. 3 (2025): EMACS
Publisher : Bina Nusantara University

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21512/emacsjournal.v7i3.13420

Abstract

This study aims to develop a hybrid Convolutional Neural Network (CNN) model for classifying the roasting levels of Coffea arabica beans by integrating RGB color and GLCM texture features. A total of 1,600 high-resolution images were used, consisting of 1,200 training images and 400 testing images, evenly distributed across four roasting levels: Green, Light, Medium, and Dark. Local feature extraction was performed using a sliding window approach to capture fine-grained color and texture information from each image. Three model types were evaluated: a CNN with RGB-only input, a CNN with GLCM-only input, and a hybrid CNN with dual inputs. The hybrid model consistently demonstrated superior performance, achieving a validation accuracy of 99.74%, with minimal misclassification and stable convergence throughout training. Furthermore, six architectural variations of the hybrid model were tested by applying dropout and L2 regularization techniques. The model combining both dropout and L2 regularization achieved the most balanced results in terms of accuracy, generalization, and training stability. This research contributes an effective feature fusion strategy for fine-grained visual classification tasks, particularly in domains where inter-class visual differences are subtle. The proposed approach offers a cost-effective and scalable solution that is well-suited for real-time implementation in small to medium-sized coffee production facilities, and it shows strong potential for broader applications in agricultural product quality assessment.