Journal of Software Engineering and Information System (SEIS)

MODEL KLASIFIKASI JARAK MANHATTAN PADA PENGENALAN CITRA SISTEM BAHASA ISYARAT BAHASA INDONESIA

2025-06-25T23:07:58+07:00

This study aims to design and implement an image recognition system for Sistem Isyarat Bahasa Indonesia (SIBI) by applying the Manhattan distance classification method. Sign language serves as a vital means of visual communication for individuals with hearing impairments and disabilities. However, public understanding of this language remains limited, often leading to ineffective communication between hearing and non-hearing communities. Therefore, an assistive system capable of accurately recognizing sign language is highly needed. The Manhattan method was selected due to its simplicity and efficiency in calculating distances between data points. The dataset used in this study was obtained from the Kaggle website, consisting of 130 training images and 130 testing images, each representing 26 alphabet letters in the SIBI system. All images underwent initial preprocessing using Jupyter Notebook, including resizing, background removal, and conversion to grayscale to facilitate feature extraction. The grayscale images were then transformed into histograms and normalized to maintain a consistent value scale. The classification process was carried out by computing the Manhattan distance between the test and training image histograms. The system was developed using MATLAB R2015a, featuring a user interface that displays classification results directly. The test results showed that out of 130 test images, 104 were accurately recognized, achieving an accuracy rate of 80%. These findings indicate that the Manhattan method is effective for use in image-based sign language recognition systems. The developed system is expected to serve as an inclusive and educational tool to enhance communication between the hearing-impaired community and the general public. Further development may involve integrating additional methods and expanding the dataset.

ANALISIS DAN VISUALISASI DATA SAMSUNG SALES MENGGUNAKAN EXPLORATORY DATA ANALYSIS PADA TABLEAU

2025-07-14T14:23:31+07:00

The development of 5G technology has a significant impact on the mobile device industry, but its adoption is uneven in many regions. The study was conducted to analyze Samsung’s 5G device sales trends globally and examine the relationship between network infrastructure, consumer preferences and device sales performance. The method used is Exploratory Data Analysis (EDA) with the help of interactive visualization through Tableau. Secondary data is obtained from Kaggle and covers the period 2019–2024, with variables such as number of units sold, network coverage, 5G average speed, and preference score. Results show that about 65% of sales come from high preference models, and since 2021 5G devices have mastered more than 70% of the global market. In addition, the Galaxy S Series model recorded preference score above 85%, showing that consumer perception is highly influential on sales performance. Visualization in the form of dashboards supports strategic understanding of markets based on regions, products, and time. This EDA-based visualization is able to provide deep insight for policymakers and manufacturers in strategizing 5G market penetration strategies more effectively and sustainably

PEMODELAN RFM & K-MEANS CLUSTERING UNTUK SEGMENTASI PELANGGAN DALAM PENJUALAN ONLINE

2025-07-14T14:14:49+07:00

The exponential growth of e-commerce platforms necessitates sophisticated customer analytics to maintain competitive advantage and optimize revenue streams. This study addresses the critical challenge of understanding heterogeneous customer purchasing behaviors in online retail environments through advanced data mining techniques. The research implements RFM (Recency, Frequency, Monetary) modeling integrated with K-Means clustering algorithm to achieve comprehensive customer segmentation for strategic marketing optimization. A quantitative-exploratory methodology was employed, utilizing a comprehensive online sales dataset comprising over 40,000 transactional records. The analytical framework involved systematic data preprocessing using Python libraries (Pandas, NumPy), followed by RFM parameter calculation and standardization through StandardScaler normalization. K-Means clustering was subsequently applied with optimal cluster determination via Elbow Method validation, yielding three distinct customer segments. Visualization and interpretation were conducted using Tableau, Matplotlib, and Seaborn for comprehensive segment characterization. Results demonstrate successful identification of strategically significant customer clusters: high-value loyal customers, moderate-engagement prospects, and potential churn-risk segments, each exhibiting distinctive RFM behavioral patterns. The segmentation framework enables targeted marketing strategy formulation, personalized customer retention programs, and optimized resource allocation. This research contributes valuable insights for e-commerce practitioners seeking data-driven approaches to enhance customer relationship management and sustain long-term business profitability in competitive online marketplaces.

PEMODELAN MACHINE LEARNING DENGAN ALGORITMA RANDOM FOREST DALAM MEMPREDIKSI RISIKO STROKE

2025-07-14T14:10:03+07:00

Stroke is one of the diseases that significantly affects health and economy, becoming the second most common cause of death in the world after coronary heart disease. Based on data from the World Health Organization (WHO), stroke is ranked second as the leading cause of death in the world after ischemic heart disease. In 2019, stroke was responsible for around 11% of total global deaths. One important way to reduce the death rate from stroke is to make prevention efforts through early prediction. Machine learning methods, especially Random Forest, are used in this study to predict the risk of stroke. The data used comes from a public dataset that includes age, gender, blood pressure, blood sugar, smoking status, and other medical history. The research process includes data pre-processing stages (data cleaning, outlier handling, and category coding), model training using the Random Forest algorithm, and model evaluation using a confusion matrix to evaluate accuracy, precision, recall, and F1 score. The evaluation results show an accuracy value of 97.55%, which indicates very good predictive performance so that this model has very good predictive performance.

KLASIFIKASI MAKANAN BERDASARKAN NILAI GIZI MENGGUNAKAN ALGORITMA RANDOM FOREST DAN TEKNIK SMOTE

2025-07-17T17:39:40+07:00

Classifying food based on nutritional content is essential for developing personalized dietary recommendation systems and promoting healthier eating habits. This study aims to construct a food classification model using the Random Forest algorithm combined with the Synthetic Minority Oversampling Technique (SMOTE) to address class imbalance in the dataset. The dataset includes various nutritional attributes such as calories, protein, fat, carbohydrates, fiber, sugar, sodium, and cholesterol, along with additional information such as food category and meal time. After preprocessing, the data were split into training and testing sets, with SMOTE applied to the training data to improve class representation. The model was trained using Random Forest and evaluated using accuracy, precision, recall, and F1-score. The results show that the model achieved an accuracy of 83.35% and an average F1-score above 0.80, with the best performance observed in majority classes. The confusion matrix analysis indicates that most predictions were accurate, although misclassifications occurred among classes with overlapping nutritional values. Protein, calories, and carbohydrates were identified as the most influential features in the classification process. These results show that combining Random Forest and SMOTE works well for creating food classification systems using nutritional data and could be useful in apps for diet recommendations and managing nutrition.

A SYSTEMS ENGINEERING APPROACH TO SUSTAINABILITY DECISION SUPPORT SYSTEM BASED ON HESI

2025-08-16T14:50:15+07:00

Climate change and global sustainability demands are prompting universities to integrate sustainable principles into all academic and non-academic activities. This research led to the design and implementation of a Sustainability Decision Support System (SDSS) at Institut Teknologi Bisnis dan Bahasa Dian Cipta Cendikia (ITBA-DCC), using the System Engineering Process (SEP) and the Higher Education Sustainability Initiative (HESI) framework. The SDSS uses Multi-Criteria Decision Making (MCDM) and interactive data visualization to assess sustainability across four dimensions: environmental, social, economic, and governance. Initial pilot testing showed an overall sustainability index of 74/100, with the highest score in the environmental dimension (82/100) and the lowest in governance (68/100). User Acceptance Testing (UAT) reported a 91% satisfaction rate among staff, lecturers, and campus leaders. Although the system relies on annual data and some manual inputs, the SDSS shows potential for adoption by similar institutions and supports Sustainable Development Goals (SDGs) in higher education.

EKSPLORASI SENTIMEN PENGGUNA X TERHADAP ISU KESEHATAN MENTAL BERBASIS MACHINE LEARNING

2025-08-16T14:50:37+07:00

Mental health has become an increasingly relevant topic in the digital era, particularly on social media platforms such as X, which serve as public spaces for expressing opinions and sharing personal experiences. This study aims to analyze public sentiment toward mental health topics on Twitter using the Multinomial Naive Bayes algorithm. Data were collected from tweets containing mental health-related keywords and processed through text cleaning and feature extraction using the TF-IDF method. The classification results showed that the model achieved an accuracy of 71%, with stronger performance in identifying negative sentiment compared to positive sentiment. A WordCloud visualization also revealed the frequent appearance of terms such as “mental,” “health,” “self,” and “disorder,” reflecting the main focus of online discussions. These findings indicate that machine learning-based sentiment analysis is effective in capturing public perceptions of mental health issues on social media. This research is expected to contribute to the development of digital communication strategies and real-time monitoring of psychosocial issues in online spaces.

OPTIMASI JARINGAN DENGAN VIRTUAL LAN (VLAN) UNTUK MENINGKATKAN EFISIENSI DATA TRANSFER

2025-08-19T16:16:29+07:00

Optimal network management is crucial in the digital era, where efficient data exchange is essential, especially in environments with numerous devices such as educational institutions and offices. The aim of this study is to evaluate the impact of implementing Virtual Local Area Networks VLAN on data transmission efficiency in computer networks. This research adopts an experimental quantitative approach through the simulation of two scenarios: networks without Virtual Local Area Networks VLAN and networks with VLAN, utilizing analytical tools such as Ping, iPerf, and Wireshark. The results show that implementing Virtual Local Area Networks VLAN can reduce average latency from 6.5 ms to 3.2 ms and increase throughput from 35 Mbps to 48 Mbps. Additionally, Virtual Local Area Networks VLAN significantly reduce broadcast traffic and packet collisions while enhancing data security through logical segmentation between departments. Therefore, Virtual Local Area Networks VLAN implementation greatly improves overall data communication efficiency and network performance.

STRATEGI PENERAPAN WIRELESS MESH NETWORK UNTUK PENGINGKATAN DAN KEANDALAN JARINGAN NIRKABEL

2025-08-20T08:16:08+07:00

The need for reliable communication networks that cover a wide area is becoming increasingly important, especially in areas with limited conventional network infrastructure. One promising solution to overcome this challenge is the implementation of Wireless Mesh Network (WMN), which offers flexibility, scalability, and redundancy through direct connectivity between nodes. This study aims to design and evaluate a Wireless Mesh Network (WMN) implementation strategy to improve the reliability and performance of wireless networks. The methods used include mesh topology analysis, adaptive routing protocol selection, and network performance simulation using throughput, delay, and packet delivery ratio parameters. The preliminary results show that the implementation of Wireless Mesh Network (WMN) with proactive routing protocols such as Optimized Link State Routing (OLSR) and a hybrid approach provides an increase in throughput of up to 35% and reduces the average delay by 20% compared to traditional infrastructure-based networks. In addition, network reliability is increased through self-healing capabilities, namely the ability of the network to reset the communication path when there is a disruption to one of the nodes. These findings indicate that the right Wireless Mesh Network (WMN) implementation strategy can significantly improve the efficiency and reliability of wireless networks, especially in environments with physical or geographical limitations. This study recommends the use of adaptive mesh topology to support continuous connectivity and more resilient network management.

PERANCANGAN SISTEM INFORMASI PEMESANAN TIKET TRAVEL BERBASIS WEB DI PT YOSSY MANDIRI

2025-08-20T09:59:45+07:00

This research aims to design a web-based ticket booking information system that can enhance the efficiency and quality of services at PT. Yossy Mandiri. This system is designed to facilitate customers in making ticket reservations online and assist the admin in managing booking data, departure schedules, and customer information. The system development method used is the waterfall method, which includes requirements analysis, system design, implementation, testing, and maintenance. The result of this research is a web-based travel ticket booking information system that can improve efficiency and service quality at PT. Yossy Mandiri. This system is expected to be a solution for the company in enhancing customer service and optimizing internal business processes.

KLASIFIKASI BUAH JERUK LEMON BERDASARKAN TINGKAT KEMATANGAN MENGGUNAKAN METODE SVM DAN NAIVE BAYES

2025-08-20T11:23:54+07:00

This study aims to develop a classification model for determining the ripeness level of lemons (Citrus limon) using digital image analysis. Two methods, namely Support Vector Machine (SVM) and Naïve Bayes Classifier (NBC), were compared to evaluate their performance in terms of accuracy and prediction consistency. The results show that SVM outperformed NBC with an accuracy of 97%, along with precision, recall, and F1-Score of 97% each. The model consistently determined lemon ripeness levels in percentage terms, such as 85% or 95%. In contrast, NBC achieved an accuracy of 82%, with precision, recall, and F1-Score of 83%, 82%, and 83%, respectively. However, NBC was more prone to classification errors, especially in distinguishing between ripe and unripe lemons. In conclusion, the SVM method proved superior to NBC in determining lemon ripeness levels, particularly in handling complex data. SVM's ability to provide accurate and consistent predictions makes it a more effective choice for helping farmers optimize the quality and quantity of lemon production. This study contributes significantly to the application of image processing technology in the agricultural sector.

PERANCANGAN SISTEM INFORMASI AKADEMIK BERBASIS WEB PADA MTSN 5 MUARO JAMBI

2025-08-20T12:29:47+07:00

The manual management of academic data has the potential to cause various obstacles, such as delays in grade distribution, data duplication, and input errors. This study aims to design a web-based academic information system at MTsN 5 Muaro Jambi as a solution to these problems. The system was developed using the Waterfall method with the PHP programming language and a MySQL database. The database design was carried out in a structured manner to support the integrity and efficiency of academic information management. The novelty of this study lies in the integration of grade management, attendance, and schedule features into a single web-based platform, which has never been implemented before at MTsN 5 Muaro Jambi. The system was designed to involve three main actors: the admin as data manager, teachers as data managers, and students as recipients of online academic information. The implementation results showed that the system was able to improve recording accuracy, accelerate information distribution, and support academic data transparency. Testing using the blackbox testing method proved that the system's main functions ran as needed, while user acceptance trials involving teachers, students, and administrative staff showed the system was easy to use and useful in supporting the academic process.