Engineering, MAthematics and Computer Science Journal (EMACS)

Leveraging Support Vector Machines and Ensemble Learning for Early Diabetes Risk Assessment: A Comparative Study

Hafizh Ash Shiddiqi — 2025-01-31

Currently, diabetes is a hidden, serious threat to human lifestyles through daily food and drink, which has become a formidable global health challenge. As a contribution, this study suggests a way to use machine learning to find people with diabetes by looking at certain health parameters. It does this by using different Support Vector Machine (SVM)-based models, such as different SVMs with different kernels, such as linear, polynomial, radial basis function, and sigmoid kernels; different ensemble bagging with SVM; and different ensemble stacking with various SVM models. The findings demonstrated that utilizing a single SVM model with a linear kernel, ensemble bagging with a linear SVM, and ensemble stacking with different SVM models yielded the most accurate results, achieving 95% accuracy in both diabetes presence and absence. This lends credence to the idea that the incorporation of a linear kernel has the potential to improve the accuracy of determining whether or not diabetic illness is present.

Parking System Application Using a Greedy Algorithm Approach

Hanis Amalia Saputri — 2025-01-31

Indonesia has recently witnessed a significant increase in the number of automobiles, reaching an estimated 17.2 million units by the end of 2022, according to the Central Statistics Agency (BPS). Extensive ownership and usage of vehicles in public parking areas, including campuses, have created a high demand for parking spaces. However, challenges still exist within the parking system, such as longer search times for available parking spaces and the lack of technological regulation, leading to uncertainty. Our research focuses on addressing these issues by employing a priority-based greedy algorithm for the nearest lift, prioritizing convenience and speed. We utilize an SQL database to store parking data, leveraging its comprehensive features for efficient processing. The result of this research is a website where customers can input their license plate numbers, processed by our algorithm to generate parking tickets, granting access to designated parking areas. The algorithm works by providing parking slot locations from even-numbered floors first; when all even-numbered floors are filled, it will then allocate parking slots on odd numbered floors. The implementation of the greedy algorithm and SQL database has proven to be efficient in the context of the nearest lift in the Binus parking lot, handling a manageable amount of data and prioritizing data processing speed over achieving the optimal solution in all scenarios

Optimizer Comparison In Convolutional Neural Network For Real Time Face Recognition

Elbert — 2025-01-31

Face recognition is one of the computer vision technologies that's used in many industries. Face recognition always used in various sector that require the verification of an individual identity. There are many ways that can be used to develop face recognition, one of them is convolutional neural network. Convolutional neural network (CNN) is a deep learning neural network that is created specifically to process and analyze visual data, such as images and videos. CNN have the ability to learn many features from visual data, making them highly effective for tasks like face recognition. There are many factors that can affect CNN performance including the optimizers that are used in the neural network. Optimizers are the algorithm that adjust weights of the neural network to minimize error between the predicted output and actual target. This study used 10 different subjects for face recognition. In this study, the CNN model uses a training algorithm called backpropagation then will compare 3 different types of optimizers. The optimizers that used in this study are Adaptive Momentum (Adam), Root Mean Square Propagation (RMSProp), and Stochastic Gradient Descent (SGD). The results of the comparison will be shown in the form of performance metrics. The performance metrics include correct classification rate (CCR) as well as the confusion matrix of each model. CNN model with SGD optimizers has the highest CCR of 97.07%.

Unlocking Pharma Market Segmentation for Strategic Growth Through Advanced Data Intelligence

Alfi Purwaningrum — 2025-01-31

Business competition compels companies to understand customer characteristics in order to maintain and enhance their competitiveness, especially in the pharmaceutical industry, which involves various customer segments such as hospitals, pharmacies, patients, and end consumers with diverse needs. Customer segmentation becomes crucial in developing effective strategies, with K-Means algorithm being one of the commonly used methods due to its simplicity and efficiency in clustering large datasets. This study combines the K-Means Clustering algorithm with the elbow method to determine the optimal number of clusters in segmenting the customer profiles of a pharmaceutical company. The analysis results reveal two main clusters: the first cluster is dominated by hospitals with higher medication purchase volumes and longer delivery distances, ranging from 8 to 131 km, while the second cluster is dominated by pharmacies with smaller purchase volumes and shorter delivery distances. These findings enable the pharmaceutical company to better understand customer characteristics and design more effective strategies to compete in the market. It is recommended that the company adjusts its marketing strategies and products based on the needs of each cluster, enhances customer relationships through loyalty programs, and optimizes distribution routes to improve operational efficiency.

Evaluation Operational of Reduce Reuse Recycle Waste Treatment Facility (TPS 3R) in Bandung City (Case Study: TPS 3R Saling Asih and TPS 3R Hikmah)

Arindita Dessi Permatasari — 2025-01-31

Waste is a problem faced by the city where currently waste is still being transported to landfill without any waste processing first. Currently, Bandung City has a waste management problem because it does not have a final disposal site (TPA) especially for waste reduction at the source and currently the waste is processed at the Sarimukti Regional Landfill, West Bandung Regency. Currently, the achievement of waste reduction in Bandung City reached 14.46% (SIPSN, 2023). The purpose of this study is to determine the operational performance of TPS3R in Bandung City as a waste treatment facility. This research was conducted at two TPS 3R in Bandung City, namely TPS 3R Saling Asih II, Maleer Village and TPS 3R Hikmah, Panjunan Village. The method in this study uses quantitative through data collection carried out on the operational conditions of TPS 3R management of 5 aspects waste management in accordance with the. Technical Guidelines for the Implementation of Reduce Reuse Recycle Waste Treatment Facility Activities (2021). From the results of this study it is known that the operational status of the management of TPS 3R Saling Asih II is in very good condition and TPS 3R Hikmah is in good condition. Optimization of TPS 3R can be done through the formation of competent community groups through various training and monitoring of operational performance of TPS 3R.

Effectiveness Analysis of RoBERTa and DistilBERT in Emotion Classification Task on Social Media Text Data

Ghinaa Zain Nabiilah — 2025-01-31

The development of social media provides various benefits in various ways, especially in the dissemination of information and communication. Through social media, users can express their opinions, or even their feelings. In this regard, sometimes users also convey information or opinions according to the user's feelings or emotions. This triggers the impact of aggressive online behavior, including cyberbullying, which triggers unhealthy debates on social media. The development of deep learning models has also been developed in several ways, especially emotion classification. In addition to using deep learning models, the development of classification tasks has also been carried out using transformer architectures, such as BERT. The development of the BERT model continues to be carried out, so this study will analyze and explore the application of BERT model development, such as RoBERTa and DistilBERT. The optimal result of this study is with an accuracy value of 92.69% using the RoBERTa model.

Developing Algorithm of Music Concepts and Operations Using The Modular Arithmetic

Kelvin Minor — 2025-01-31

The rapid development of digital music technology is closely intertwined with advancements in both music theory and mathematical formalism. This study aims to bridge the gap between these fields by exploring how mathematical concepts can enhance the understanding and analysis of music theory. Specifically, the research focuses on the application of modular arithmetic to analyze the circular structure of the chromatic scale, a key concept in music. Modular arithmetic enables the identification of patterns in pitch relationships and the manipulation of musical elements like transposition and interval calculations. In addition to modular arithmetic, the study also highlights the role of regular expressions in music theory. Regular expressions provide powerful tools for pattern matching, which can be applied to recognize and categorize musical components, such as enharmonic equivalents (notes that sound the same but are named differently). These tools allow for the development of algorithms capable of generating chords from given notes or identifying chords from existing sets of notes. By integrating modular arithmetic and regular expressions, the study proposes a framework for developing mathematical models and algorithms to facilitate digital music analysis. This approach not only enhances the theoretical understanding of music but also holds practical applications in digital music production and education.

Indoor Positioning System using Gaussian Mixture Model on BLE Fingerprint

Maximilianus Maria Kolbe Lie — 2025-01-31

After the release of Bluetooth Low Energy (BLE), people have been trying to use Bluetooth as an alternative source to solve indoor positioning. Unfortunately, due to the nature of Bluetooth about proximity, the propagated signal is very fluctuating. This decreases the accuracy considerably and has become one of the main problems in using Bluetooth. To combat the signal fluctuations, we propose a fingerprinting-based concept of using received signal strength (RSS) frequency distribution values as the data in the radio map, which is termed Frequency Distribution Radio Map (FDRM). We also propose a probabilistic fingerprinting-based algorithm utilizing FDRM using Gaussian Mixture Model (GMM) as the probability distribution function (PDF). In the offline phase, we compare 2 methods: k-Means only, and k-Means with Expectation-Maximization (EM); to learn the FDRM. This resulting a probability distribution function (PDF) of the RSS in each reference points for each BLEs. In the online phase, k-Nearest Neighbour (KNN) and weighted average are used to estimate the receiver’s location. The proposed method is validated over 3 different datasets taken from a 4 m x 6 m classroom equipped with chairs and tables. The experiment shows that the proposed fingerprint and model are better in capturing the environment, including the signal fluctuation. By using only k-Means in obtaining the GMM, it achieved mean error of 98.18 cm and standard deviation of 56.11 cm. By adding EM, there will be a trade-off between mean error with better standard deviation and lower computing time. It achieved standard deviation of 47.99 cm and mean error of 112.24 cm.

Cyber Security Awareness Simulation for Web Phishing in E-Commerce

Jason Matthew Sutanto — 2025-01-31

The development of information and communication technology has changed people's behavior in conducting transactions for buying and selling goods and services. With e-commerce, people can conduct transactions for buying and selling goods or services digitally. Although e-commerce can provide several positive impacts for society, one of which is through the ease of buying and selling goods digitally, e-commerce also has negative impacts for society, one of which is the risk of cybercrime, such as hacking, theft, and fraud. Therefore, consumers as e-commerce actors need to be protected, guarded, and secured. Protection in cyberspace is not only about strengthening the existing security system, but e-commerce users also need to be given an understanding and knowledge about cybersecurity to prevent data theft and hacking through phishing or other social engineering attacks. This study aims to build an application or feature that can increase security awareness for e-commerce users. The method used is the Waterfall method, as well as SIT (System Integration Testing) and UAT (User Acceptance Testing). Algorithm design involves the Dart programming language and the Flutter framework. The results of the study show that the applications or features created can increase security awareness among users and prevent the risk of cybercrime threats, such as fraud or hacking.

Classifying Viral Twitter with Transformer Models and Multi-Layer Perceptron

Jeffrey Junior Tedjasulaksana — 2025-01-31

The classification of virality levels in Indonesian tweets is explored in this research using advanced natural language processing techniques and machine learning algorithms. Transformer models such as RoBERTa for sentiment analysis and XLNet for text embedding, alongside Multi-Layer Perceptron (MLP) classifiers, are leveraged to address the challenge of predicting tweet virality. Emotion features are incorporated, and cost-sensitive methods for handling class imbalance are implemented, resulting in robust performance demonstrated by our model. Intriguing correlations between tweet sentiment, emotion distribution, and virality levels are uncovered through sentiment analysis and emotion detection. The efficacy of XLNet in capturing contextual nuances, outperforming BERTweet, is highlighted by our findings. Furthermore, the integration of emotion features and cost-sensitive methods enhances the model's predictive accuracy, offering valuable insights for marketers and businesses seeking to optimize their social media strategies. The proposed model achieves an accuracy of 95% and an F1-Score of 59%.

Digital Readiness Assesment to Improve Quality of Life Using SWOT Analysis: Case Study for Natuna Regency Area

Irma Rizkia — 2025-01-31

Natuna Regency, as an outlying region, faces challenges in enhancing its digital readiness to effectively implement digital transformation. Currently, the digital readiness of the region remains at the survival level. This research aims to explore the most effective strategies in improving the digital readiness of the city government, and its impacts on public services and community participation. An empirical approach is employed through a case study in Natuna Regency. The research begins with a literature review and a pilot case study on the digital readiness of a region, validated through questionnaires based on the Garuda Digital Transformation Framework (GDTF) developed by ITB as the basis for measuring digital readiness in Natuna Regency. SWOT analysis is utilized to provide strategies for Natuna Regency. The digital transformation readiness of Natuna Regency is at the survival level, with a score of 60.37. Tangible impacts of digital transformation are evident in the business processes conducted by government agencies and the ecosystem of governance and society in Natuna Regency. The primary recommendation is to leverage internal strengths to address external challenges and capitalize on opportunities in the digital transformation era. Strategies include leveraging high digital literacy, understanding data security, developing technological infrastructure, enhancing human resource skills, and collaborating with the private sector and educational institutions. By integrating these strategies, Natuna Regency can more effectively respond to technological challenges and leverage digital transformation opportunities for sustainable progress.

Research on The Empirical Analysis of Bitcoin and Gasoline Return

Asysta Amalia Pasaribu — 2025-01-31

Investment is an activity that is popular nowadays. Profitable investments are the hope of every investor. By investing. investors expect the invested assets to generate returns and to obtain profits for future life In investment studies. the most frequently discussed topic is the fluctuations. whether increases or decreases. of an asset's price (stocks). The risk of investment is loss in financial. The fluctuations of stock prices represent risks in the investment field. One measure used to determine gains and losses from stock prices is the return. To know return from data. we may use the compound return formula. Returns have empirical facts that require several tests. In this study. the empirical facts of returns are that the returns are not autocorrelated (autocorrelation function) and that the returns are leptokurtic distributed (thick-tailed distribution). We use the price data of Bitcoin (BTC) and Gasoline (UGA) from January 1. 2019. to December 31. 2023. The main of purpose of this research is to show empirical analysis of the Bitcoin and Gasoline return data. The results of the empirical analysis show that the return of stock price for Bitcoin (BTC) and Gasoline (UGA) meet the empirical properties of returns so that they can capture a good volatility model.

A Horizontally Scalable WebSocket Architecture for Cost-Effective Online Examination Proctoring System on AWS Cloud Infrastructure

Eko Cahyo Nugroho — 2025-01-31

In this research work we present the cost-effective prototype of a WebSocket server with a horizontal scaling feature on AWS Cloud Service. AWS API Gateway for establishing WebSocket connections also works but is exceedingly expensive for schools. The solution presented in this study proposes an on-premise WebSocket server deployed at AWS EC2 instances. The server utilizes Node. js's cluster module to make the most out of the CPU's cores and has also implemented a Redis pub/sub mechanism to easily horizontal scale it to many EC2 instances. The system architecture utilizes DynamoDB to store students' proctoring status recorded on the first attempt at the quiz. Then, the real status update is delivered by WebSocket message. The implementation shows effective real-time monitoring capabilities for online examinations, including student activity tracking, automated disconnection detection, and proctor-student interaction features. The results show improved cost efficiency compared to API Gateway as the WebSocket server. This solution provides schools with a cost-effective and reliable proctoring feature in LMS for implementing online examination proctoring systems at scale.

A Analysis and design of Employee Data Processing Application Using QR Code With Webcam (Case Study: PT. Adhicon Perkasa)

Ayuliana — 2025-01-31

Employee attendance is a process of recording employee attendance that is commonly used in companies using technology such as fingerprint attendance, web-based attendance, android-based attendance, Quick Response Code attendance and others. The use of technology is expected to help in managing attendance data. The purpose of this study is to design a Web and Android-Based Quick Response Code Attendance Application that can facilitate the process of attendance and daily attendance recapitulation so that the results obtained are more accurate, then the stored data is used to calculate employee meal money needs per day. The research method is divided into problem identification, solution development, and system design. The interview approach by taking a case study at PT. Adhicon Perkasa. The results of this study are the Quick Response Code Attendance Application that can make the attendance process more accurate, minimize human error, and facilitate attendance recapitulation for calculating daily meal money. It is concluded that with this application, employee performance is expected to increase because there is no need to make manual attendance reports and the calculation of meal money is more accurate.

Origin of Trigonometry Formula cos (A+B), sin (A+B), cos (A-B), sin (A-B) using the Euler Formula

Wikaria Gazali — 2025-01-31

This paper discusses the Origin of Trigonometric Formulas cos(A+B), sin(A+B), cos(A-B), and sin(A-B). To obtain the model, the writer uses Euler's formula e^iθ = cos⁡ θ + i .sin " θ. In this derivation we will get the formulas sin (A+B), cos (A+B), sin (A-B), and cos (A-B).

Editorial Page and Table of Content

Alexander Agung Santoso Gunawan — 2025-01-31