Jurnal CoSciTech (Computer Science and Information Technology)

Perancangan Aplikasi Tata Usaha Kayu Dengan Traceability System Berbasis Web Pada PT Bumi Trikama Jayasri

Irma Aprilia — 2025-12-13

PT Bumi Trikama Jayasri is a timber management company that faces challenges in recording timber distribution, such as data inaccuracies, delays in real-time input, and a lack of transparency in reporting. This research aims to design a web-based timber administration application with a traceability system that facilitates efficient recording, tracking, and reporting of timber distribution. The system development method uses the Waterfall model, consisting of analysis, design, implementation, testing, and maintenance stages. The technologies used include PHP with the Laravel framework and a MySQL database. The system is equipped with tracking features for distribution from TPN (Timber Collection Point) to barges, as well as reporting and data history modules. The Requirement Traceability Matrix (RTM) is used to verify the fulfillment of system requirements. Test results show that the system improves the efficiency and accuracy of record-keeping and supports faster and more accurate decision-making. This system is expected to serve as an effective digital solution for timber distribution management in the company.

Waktu Respons Transmisi Data Dalam Impelentasi Algoritma A-Star Pada Sistem Pengambilan Tempat Sampah IoT Berbasis Telegram

Sasridarti Tari Sejahtera Sabuna — 2025-12-13

Effective waste management requires an efficient monitoring system and collection route determination. One of the main challenges is the delay in delivering information that is not yet integrated into an automated schedule. To address this, this research designs an IoT-based waste bin monitoring and collection route determination system using the A-Star algorithm, as this algorithm can efficiently determine routes by considering the distance between nodes as well as the waste volume conditions. The system is equipped with a website that displays the waste volume conditions at each node, with the resulting routes sent to officers via the Telegram platform. The focus of this research is to analyze the system's performance, specifically the data transmission response time, which is defined as the time span starting from when the sensor detects the waste volume, data is sent by the ESP32 to the server, the server processes the algorithm, until the route information or notification is received by the user. The results show that the average processing time of the A-Star algorithm on the server is 0.0135 s, the average data transmission delay is 1.2659 s or 1265.9 ms, which is categorized as POOR based on the TIPHON standard, and the average system update time is 6.38473 s.

Design and Development of Web-based Inventory System with Sales Prediction Using Time Series Forecasting

Ridwan Rahmat — 2025-12-13

Effective inventory management is a crucial aspect of company operations to predict future stock requirements and product demand. This research aims to design and develop a web-based inventory system with sales prediction using Time Series Forecasting algorithms at CV Adio Loop Engineering. The development method used is waterfall with Long Short-Term Memory (LSTM) approach as a prediction model based on historical inventory transaction data. The system has comprehensive features including dashboard with information on total products, purchases, sales, categories, and suppliers; prediction module for selecting products and prediction types (demand/stock) with time estimation; master data for managing categories, products, and suppliers; transaction modules for purchasing, sales, and inventory; stock movement; low stock alerts; inventory reports; and human resource management with login/logout security system. All modules are equipped with complete CRUD functions. Test results show that the system is capable of providing accurate predictions and improving operational efficiency in inventory management and future stock requirement planning.

Designing a Cashier Website for Warkop Disini Aja Using the Laravel Framework with UAT Testing and Usability Testing Checklist

Taufik Rahman — 2025-12-14

Advances in information technology have a significant influence on various sectors, including Micro, Small, and Medium Enterprises (MSMEs). However, most MSMEs still have not adopted digital systems in their operational activities. One example is Warkop Diini Aja, which until now still applies manual recording to the cashier transaction process. The use of the conventional system poses various obstacles, such as delays in making reports, potential for recording errors, and difficulties in monitoring sales history. Based on these problems, this research aims to design a web-based cashier system that is able to replace manual methods to be more efficient and structured. System development is carried out using the Laravel framework with a waterfall software development model, which includes the stages of needs analysis, design, implementation, testing, and maintenance. Research data was obtained through direct observation of operational processes, interviews with business owners, and literature review to strengthen the theoretical basis. The developed system has two main roles, namely admin and cashier, which are equipped with menu management features, sales transactions, export reports in digital format, and transaction history monitoring. Based on the results of the User Acceptance Testing (UAT) test, all system functions are declared to run according to user needs. In addition, the results of the Usability testing Checklist show a user satisfaction rate of 95%, which is classified as very feasible. Thus, this website-based cashier system has been proven to be able to improve operational efficiency, make it easier to record transactions, strengthen financial report transparency, and support more modern and digital business management.

Design and Development of a Web-Based Employee Promotion Assessment System at PT Timbul Mandiri Agung Using the Weighted Scoring Method

Iklil Fakhriyatin — 2025-12-14

This study aims to design and develop a web-based employee promotion evaluation system at PT Timbul Mandiri Agung using the Weighted Scoring method. The system was developed to improve efficiency, accuracy, and transparency in the evaluation process, which was previously conducted manually. The Weighted Scoring method enables objective decision-making by assigning weights to evaluation criteria such as performance, work experience, skills, attitude, discipline, and attendance. The study adopted the Waterfall model for system development, including requirements analysis, system design, implementation, testing, and evaluation. The results indicate that the proposed system enhances evaluation accuracy, accelerates decision-making processes, and increases employee trust in the company’s evaluation system.

Comparison of random forest and xgboost algorithms in credit card fraud classification

Asrul Abdullah — 2025-12-14

Credit card fraud is a serious issue that can cause significant losses for both consumers and financial service providers. Therefore, a reliable and accurate fraud detection system is essential. The research adopts the CRISP-DM methodology, which includes six phases: Business Understanding, Data Understanding, Data Preparation, Modeling, Evaluation, and Deployment. The dataset used was obtained from the Kaggle platform, consisting of 1,048,574 rows and 23 Features, including transaction amount, merchant category, location, and customer attributes. Model evaluation was conducted using a Confusion Matrix with accuracy, precision, recall , and F1-score as performance metrics. The evaluation results indicate that Xgboost outperforms Random Forest, achieving an accuracy of 99.19%, precision of 98.73%, recall of 99.66%, and F1-score of 99.19%. In comparison, Random Forest achieved an accuracy of 97.68%, precision of 97.38%, recall of 98.01%, and F1-score of 97.69%. These results demonstrate that Xgboost is more effective in consistently identifying fraud ulent transactions. Furthermore, this study successfully developed a web-based application using the Streamlit framework, integrating both models interactively to allow users to input data and obtain classification results in real time. Thus, this study has successfully achieved three main objectives: identifying the most suitable algorithm for fraud classification, thoroughly evaluating model performance, and developing an application as a decision support system for credit card fraud detection.

Prediksi Harga Mobil Bekas Menggunakan Algoritma Support Vector Regression

Herlangga Herlangga — 2025-12-14

The growth of the automotive industry in Indonesia has contributed to high demand for used cars as a more economical alternative to new cars. However, determining the price of used cars is often a challenge for showrooms and prospective buyers because it involves many factors and is subjective. This study aims to develop a used car price prediction model using the Support Vector Regression (SVR) algorithm with a Radial Basis Function (RBF) kernel approach. A total of 1,000 entries were obtained through web scraping from the cintamobil.com website. The research methodology refers to the CRISP-DM framework, starting from business understanding to model deployment through a web application using Streamlit. The preprocessing process involves handling missing values, outliers, data duplication, and numerical and categorical feature transformations. The SVR model was evaluated using RMSE, MAPE, and MAE metrics to assess prediction accuracy. The results show that SVR is capable of providing fairly accurate price predictions, with parameters C=1, gamma=0.1, and epsilon=0.1 producing the best performance, namely an MAE value of IDR 6,472,572, an RMSE of IDR 8,958,555, and a MAPE of 3.41%. Referring to the prediction accuracy category based on the MAPE value, where a MAPE value ≤ 10% is categorized as high accuracy, it can be concluded that this model has high prediction accuracy. This shows that the SVR model used is capable of estimating used car prices with a low error rate and good accuracy.

Klasifikasi Citra Penyakit Daun Tomat Menggunakan Metode Convolutional Neural Network (CNN) Dengan Arsitektur VGG-19

Fitri Handayani — 2025-12-14

Tomatoes, known as Solanum lycopersicum in Latin, are a type of horticultural commodity with high economic value in Indonesia.Tomato production can decrease due to leaf diseases that are hard to identify manually because the symptoms of different diseases often appear similar. The purpose of this study is to apply a deep learning-based tomato leaf disease classification system using the Convolutional Neural Network (CNN) VGG-19 architecture. The dataset was obtained from Kaggle and contains 6,600 images of tomato leaves divided into six disease classes and one healthy leaf class. The research stages include preprocessing (resizing, normalization), data augmentation, dataset division (80% training, 20% testing), model training with transfer learning, and fine-tuning for optimization. The evaluation using the confusion matrix and classification report includes accuracy, precision, recall, and F1-score. Test results show that the VGG-19 model achieved 97% accuracy on the test data, with an average precision, recall, and F1-score of 0.97. These findings show that VGG-19 effectively identifies tomato leaf diseases and could be applied in web- or mobile-based detection systems to help farmers with early diagnosis and proper treatment.

Implementation of Ant Colony Optimization for the Shortest Route in J&T Package Delivery

Rahandya Cahyo — 2025-12-14

Parcel delivery is a logistics service that requires speed and efficiency, especially in determining delivery routes. The choice of this topic is based on the problem faced by J&T delivery in Kubu Raya, particularly Desa Kapur, where long travel distances often result in inefficiency. This study applies the Ant Colony Optimization (ACO) algorithm to identify the shortest route for parcel delivery. ACO mimics the behavior of ants in finding optimal paths based on pheromone intensity. Location data were obtained using coordinates from the Google Maps API and modeled into a weighted graph, where nodes represent delivery points and edges represent distances. The optimization process was carried out by simulating the movement of ant agents to evaluate alternative routes, followed by pheromone updates on the more efficient paths. The results indicate that ACO successfully generated more efficient delivery routes compared to conventional methods, achieving a distance reduction of 28.29%, equivalent to approximately 10.68 km saved. This efficiency contributes to reduced travel time and operational costs. The optimized routes were also visualized through an interactive map using Leaflet.js to facilitate analysis and interpretation. Therefore, ACO is proven to be effective in optimizing delivery routes and has strong potential for real-world application in courier services.

Improvement of Crescent Image Quality Based on Contrast Using the Histogram Equalization, AHE, and CLAHE Methods

Ady Suprayitno — 2025-12-14

The determination of the beginning of the Hijri month is often aided by digital imaging technology, but the quality of the crescent images produced often faces the challenge of very low contrast. The faint light of the crescent is difficult to distinguish from the still bright background of the evening sky, exacerbated by atmospheric conditions and camera sensor noise that reduce visual quality. To improve the image, many still perform manual contrast enhancement. On the other hand, the selection of contrast enhancement methods is often without a measurable basis. This study aims to conduct a comparative performance evaluation between three contrast enhancement methods: Histogram Equalization (HE), Adaptive Histogram Equalization (AHE), and Contrast Limited Adaptive Histogram Equalization (CLAHE). The goal is to identify the most suitable technique for improving the quality of crescent images, the specific application of which has not been widely explored. A total of 30 crescent images were tested through a quantitative evaluation approach using the Mean Squared Error (MSE) and Peak Signal-to-Noise Ratio (PSNR) metrics. The results show that CLAHE provides the best performance with the lowest average MSE (89.97) and the highest PSNR (30.92 dB), demonstrating the best ability to balance contrast enhancement and distortion reduction. In contrast, the HE and AHE methods produce high MSE and low PSNR values, indicating significant visual distortion. Thus, CLAHE is recommended as the most reliable method for improving the quality of crescent images based on contrast in digital technology-based observation systems. For further research, it is recommended to explore the automatic determination of CLAHE parameters and the use of additional evaluation metrics such as SSIM (Structural Similarity Index Measure).

Identification of Chili Leaf Diseases Using DenseNet169 Architecture

Putri Rizka Setiarini — 2025-12-14

Chili is a high-value agricultural commodity in Indonesia, but its production is often hindered by leaf diseases such as spots, curling, and yellowing. Early identification of these diseases is crucial to prevent significant yield losses. This study aims to develop an automated system for identifying chili leaf diseases using the DenseNet169 Deep Learning architecture, implemented via a web-based platform. The methodology includes data collection from Roboflow.com (3,610 images of chili leaves across four classes: spots, curling, yellowing, and healthy), data preprocessing, augmentation, model training, and evaluation. The results demonstrate that the DenseNet169 model achieves an accuracy of 98%, with consistent precision, recall, and *F1-score* values for each class. The model is integrated into a Flask-based web application, allowing users to upload images of chili leaves for disease prediction and treatment recommendations. This system is expected to assist farmers in early disease detection, thereby improving cultivation efficiency and reducing crop failure risks.

Analysis of Promotional Media Recommendations for Nurul Falah High School Admissions in Pekanbaru Using the K-Means Clustering Algorithm

Muhammad Taufiq Hidayat — 2025-12-14

The advancement of information technology has significantly transformed how educational institutions conduct promotional activities and student admissions. The shift toward digital behavior among society requires schools to adopt more adaptive and data-driven marketing strategies. SMA Nurul Falah Pekanbaru, as one of the private high schools, has experienced a decrease in student enrollment after the Covid-19 pandemic. Conventional promotional methods such as banners, billboards, and printed brochures have proven less effective in reaching prospective students. This research aims to analyze the effectiveness of promotional media based on student characteristics using the K-Means Clustering algorithm as a segmentation method. The dataset was obtained from three years of PPDB registration records, including demographic, socioeconomic, school origin, and promotion media information. The analysis process involved several stages, namely data preprocessing, exploratory data analysis (EDA), determination of the optimal number of clusters using the Elbow and Silhouette methods, and the development of a web-based recommendation system using Python, PHP, and MySQL. The results indicate that the optimal number of clusters is k=4 with a Silhouette Score of 0.351. The four clusters represent distinct behavioral patterns in accessing educational information, with digital media emerging as the most effective channel. The developed recommendation system provides decision support for the school in designing promotional strategies that are more efficient, measurable, and accurately targeted through data analytics-based insights.

Optimalisasi Desain Basis Data E-Commerce untuk Menjamin Integritas Data (Studi Kasus: Web E-Commerce Rie.charge)

Medhanita Dewi Renanti — 2025-12-14

E-commerce platforms require accurate data management, but often face the challenge of data redundancy that threatens system integrity. This study aims to design an optimal database for the Rie.charge e-commerce case study. The research method used is the Database Life Cycle (DBLC) approach, with a focus on logical design. This process systematically applies normalization techniques to transform data from Unnormalized Form (UNF), through First Normal Form (1NF) and Second Normal Form (2NF) , to Third Normal Form (3NF). The results show that the 3NF design was successfully achieved, as evidenced by the redundancy rate analysis, which shows a significant and effective reduction in data redundancy. This optimal design has been proven to successfully eliminate data anomalies (insertion, update, deletion) and ultimately ensure data integrity.

Implementation of Deep Learning for Disease Classification in Oil Palm Leaves Using the MobileNetV2 Architecture

Habil Putra Arianda — 2025-12-15

Accurate and efficient identification of diseases in oil palm leaves is a crucial challenge in maintaining plantation productivity and preventing significant crop losses. Limited access to experts and slow detection in the field are often obstacles. This study aims to develop a palm oil leaf disease classification model using a deep learning approach based on Convolutional Neural Network (CNN) with MobileNetV2 architecture. This model utilizes a transfer learning strategy from pre-trained ImageNet weights and is optimized through a two-phase training strategy on a dataset consisting of 1200 augmented oil palm leaf images, covering four classes, namely Healthy Sample, Fusarium Wilt, Parlatoria Blanchardi, and Rachis Blight. Model testing results show an accuracy of 85% on separate test data. The MobileNetV2 architecture was chosen for its lightweight characteristics, making this model efficient and highly suitable for implementation on mobile devices to assist in rapid disease identification in the field and support decision-making by farmers.

Design and Development of a Web-Based Point Of Sales Application with MVC Architecture Using the Laravel Framework at PT Palokoto Agro Industri

Azi As'ari — 2025-12-15

PT Palokoto Agro Industri still relies on Microsoft Excel for warehouse record-keeping, which is ineffective for managing large-scale data, prone to errors, and lacks security. The manual stock update process increases workload and reduces data accuracy. Furthermore, the absence of real-time access limits managers in monitoring warehouse activities and making timely decisions. To address these issues, this study developed a web-based Point Of Sales (POS) application. The application was built using the Model-View-Controller (MVC) architecture and the Laravel framework, equipped with features that align with warehouse recording standards, such as managerial access, automatic calculation of incoming and outgoing goods, and fast report generation. This research applied the Research and Development (R&D) method with a prototyping approach. The application was evaluated using the ISO/IEC 25010 standard, and the results showed that it fulfilled all aspects of software quality, including functional suitability, reliability, usability, performance efficiency, maintainability, portability, compatibility, and security. Therefore, the developed application meets the required quality criteria and can serve as a structured solution for warehouse record-keeping at PT Palokoto Agro Industri.

The Application Of The Customer Satisfaction Index Method In Consumer Satisfaction With Service At Toko Cikal Aquarium.

Satria Pasaribu — 2025-12-15

This study aims to determine the level of customer satisfaction with the services at Toko Cikal Aquarium by applying the
Customer Satisfaction Index (CSI) method. The store faces several issues, including a manual sales system, difficulties in
obtaining customer satisfaction data, and the lack of a digital platform to support service quality evaluation. This research
adopts a quantitative approach with data collection methods including observation, interviews, and questionnaires. The CSI
method is used to analyze key service attributes such as location, individual service, store attributes, and product quality. The
system design follows the waterfall approach, utilizing PHP for programming, MySQL for the database, and UML modeling
techniques such as use case, class, activity, and sequence diagrams. The results show that implementing the CSI method in a
web-based system enables the seller to more easily identify which attributes influence customer satisfaction and make
improvements based on the collected data. This system also facilitates documentation, evaluation, and strategic decision-making
in managing services at Toko Cikal Aquarium.

Implementation of the User Experience Questionnaire (UEQ) Method as an Evaluation of User Experience in the Kita Bestee Application

M. Anton Permana — 2025-12-15

User experience plays a crucial role in determining the success of a digital application, including the Kita Bestee application used by facilitator companions at PT Bank BTPN Syariah Tbk. This study aims to evaluate the user experience (UX) of the Kita Bestee application using the User Experience Questionnaire (UEQ), which covers six key dimensions: attractiveness, clarity, efficiency, accuracy, stimulation, and novelty. Using a quantitative approach, data were collected through questionnaires distributed to 75 active facilitators in Sukabumi City and Regency. The data were analyzed with the UEQ Data Analysis Tool version 12 and supported by validity and reliability testing using SPSS version 26. The results showed that clarity received the highest score of 0.88, followed by attractiveness (0.76) and stimulation (0.73), while efficiency and novelty scored lower and require improvement. Based on UEQ international benchmarks, most UX dimensions fall into the above average category but have not yet reached the excellent level. Overall, the findings indicate that the Kita Bestee application provides a fairly positive user experience, though enhancements are still needed, especially in terms of innovation and interface efficiency, and the study offers strategic recommendations to support future UX improvements.

Implementation of yolov8 nano in an iot-based oyster mushroom cultivation monitoring system

Andi Nopiandi — 2025-12-15

Oyster mushrooms are one of the agricultural commodities with high economic value and are widely cultivated in Indonesia. However, the conventional process of monitoring their growth is still carried out manually, which requires considerable time and labor while also being prone to errors in decision-making. To address this issue, this study developed an automatic oyster mushroom growth monitoring system using Internet of Things (IoT) and Artificial Intelligence (AI) technologies. The system uses a DHT22 sensor to measure temperature and humidity, a BH1750 sensor to measure light intensity, and an ESP32-CAM module to capture mushroom images. The data is transmitted through the ESP32 and analyzed using Python, while the images are processed by a YOLOv8 Nano model to classify mushroom growth stages into baglog, young mushrooms, and ready-to-harvest mushrooms. The monitoring results are displayed in real time on a dashboard and stored in a MySQL database. The model training results show fairly good performance, with an average precision of 0.69, recall of 0.78, and a mean Average Precision (mAP@0.5) of 0.71. Further testing was conducted on 15 test images for each mushroom stage, and all images were successfully detected according to their actual classes. Additionally, tests conducted on 10 negative images (without mushroom objects) also supported the system’s reliability. The success of the system is further supported by stable network connectivity for data transmission, adequate lighting in the cultivation room during image capture, and automatic adjustment of temperature and humidity according to the mushroom growth phase. This system demonstrates its capability to monitor mushroom growth conditions automatically and accurately, offering a practical solution for supporting more modern and efficient mushroom cultivation practices.

Penerapan cloud computing dalam meningkatkan industri kreatif di kota lamongan

Rana Atikah Ardlianti — 2025-12-18

The current pace of science and technology development is accelerating and has become a crucial aspect of societal life. Lamongan is one of the cities in East Java with a variety of scattered Creative Industries. One of the most popular Creative Industries among the public is the F&B Industry. In this study, the F&B business selected is a Lamongan Soto stall (Warung Soto Lamongan). After conducting the research, several issues were found to arise because the services regarding ordering and payment at this Soto stall are still conventional, leading to sub-optimal processes and time-consuming queues. The same also applies to the processes of payment, promotion, and even bookkeeping. The most common and easy-to-implement technology is cloud computing, as this technology is user-friendly and can enhance business efficiency. The purpose of this study is to offer a solution to Creative Industry players so they can grow and compete by implementing a technology-based platform.

Prediksi penyakit diabetes mellitus menggunakan metode case based reasoning

Auliyya Rahimah — 2025-12-26

Diabetes Mellitus (DM) is a chronic disease that can lead to serious complications if not detected and treated early. According to data from WHO and the Indonesian Ministry of Health, the prevalence of DM continues to rise each year, highlighting the need for a diagnostic support system that is both fast and accurate. This study aims to develop an expert system capable of predicting Diabetes Mellitus using the Case Based Reasoning (CBR) method. CBR is applied because it solves new problems by comparing them to previous cases based on the similarity of symptoms. The system incorporates 20 symptoms classified into two types of DM: type 1 and type 2. The prediction process follows the four main stages of CBR: retrieve, reuse, revise, and retain. Test results show that the system can predict the disease with an accuracy rate of over 90%, and user feedback through Blackbox Testing and User Acceptance Testing (UAT) confirms its usability. This expert system is expected to serve as an initial consultation tool to help users obtain early information related to potential DM quickly, easily, and efficiently.

Perbandingan Kinerja Model GARCH Dan LSTM Dalam Memprediksi Volatilitas Harian IHSG

Gabriel Sitorus — 2025-12-26

This study compares the performance of the Generalized Autoregressive Conditional Heteroskedasticity (GARCH) and Long Short-Term Memory (LSTM) models in predicting daily volatility of the Jakarta Composite Index (JCI) for the 2016–2025 period. Volatility is an important indicator in assessing market risk and uncertainty, so accurate prediction methods are needed by investors, analysts, and policymakers. The JCI closing price data is converted into log returns and processed through cleaning, normalization, and sequence formation stages for modeling purposes. The GARCH(1,1) model is used to capture the nature of volatility clustering through a conditional variance approach, while LSTM is utilized to study non-linear patterns and long-term relationships in time series. The results show that GARCH(1,1) is able to describe volatility patterns in general, but is less responsive to sudden changes in volatility. In contrast, the LSTM model provides superior prediction performance with low prediction errors and high coefficient of determination values. These findings indicate that the deep learning approach is more effective in modeling the volatility dynamics of the Jakarta Composite Index (JCI) than traditional econometric methods, especially under volatile market conditions.

Keywords: JCI Volatility, GARCH, LSTM, Time Series Forecasting, Deep Learning

Classification of DDoS attacks using the random forest method and class weight technique on the CICDDoS2019 dataset

Desti Mualfah — 2025-12-26

The rapid advancement of information technology has significantly influenced various aspects of life, including an increasing reliance on network-based services. However, this dependence has also led to the emergence of more complex cybersecurity threats, one of the most prominent being Distributed Denial of Service (DDoS) attacks. These attacks can disrupt service availability by overwhelming target systems with excessive traffic. A major challenge in detecting DDoS attacks lies in the wide variety of attack patterns and the class imbalance that commonly occurs in network traffic datasets. To address these issues, a machine learning–based approach capable of handling complex attack behaviors while compensating for imbalanced data distribution is required. One potential solution is the use of the Random Forest algorithm with class-weight techniques, applied to the CICDDoS2019 dataset. The research procedure includes data collection and exploration, preprocessing steps such as handling missing and infinite values, encoding categorical attributes, and feature normalization. The dataset is then divided into training and testing subsets before being processed by the Random Forest model. Model evaluation is conducted using a confusion matrix along with accuracy, precision, recall, and F1-score metrics. Experimental results show that incorporating class weight significantly improves model performance, achieving an accuracy of 99.98%, precision of 99.98%, recall of 99.97%, and an F1-score of 99.97%. These findings demonstrate that the proposed approach is highly effective for accurately detecting and classifying DDoS attacks.

Implementasi Sistem Informasi Berbasis Web Pada Pengelolaan Arsip Bagian Sertifikasi Balai BPOM Pekanbaru

Darmanta Sukrianto — 2025-12-26

Archives play a crucial role in the administrative and management activities of an organization as they contain information about daily operations. The BBPOM (Food and Drug Supervisory Agency) in Pekanbaru is a technical unit responsible for overseeing food and drug products in accordance with legal regulations. One of the divisions at BBPOM Pekanbaru is the Certification Division, which is tasked with evaluating the production facilities of food and drug products. Currently, the Certification Division still uses a simple archiving management system, where all physical documents are scanned and stored in Google Drive, while the physical documents are kept in filing cabinets. Although the arrangement is neat, there is still a higher risk of data damage or loss, and it also requires a large storage space. Developing an information system for efficient archiving takes time, but by using the Rapid Application Development (RAD) method, the system can be completed quickly. This web-based archiving information system is designed to make data searches faster by using a search feature for finding document numbers, names, and types. The system also facilitates monitoring of storage spaces and maintaining archived documents, while minimizing the risk of document damage or loss, as access to the archiving system requires a username and password for login.

Comparison of SARIMA and Prophet models for forecasting international tourist arrivals to Indonesia based on monthly time series data

M Ilmi Alfaridzi — 2025-12-26

Forecasting international tourist arrivals is a critical aspect of tourism planning and policy-making. This study compares two time series forecasting methods, Seasonal Autoregressive Integrated Moving Average (SARIMA) and Prophet in modeling and predicting the monthly number of international tourists visiting Indonesia, based on data from January 2018 to May 2025. The methodology includes data preprocessing, stationarity testing using the Augmented Dickey-Fuller test, and selecting optimal SARIMA parameters based on the lowest AIC. Model performance was evaluated using MAE and RMSE on the testing data from January to May 2025. The results indicate that SARIMA outperforms Prophet, with a lower average MAE of 1336.41 and RMSE of 1616.67, compared to Prophet’s MAE of 5591.33 and RMSE of 5739.71. Based on this evaluation, SARIMA was selected as the best model and used to project international tourist visits for the period June to December 2025. These findings highlight SARIMA’s superior ability to capture seasonal patterns in tourism data, making it a reliable tool for short-term tourism forecasting in Indonesia.

Deteksi Spam Email Multibahasa: Menggunakan Cross-lingual Transfer Learning

Galih Mahalisa — 2025-12-26

Targeting the challenge of text classification in Indonesian, which often faces a scarcity of adequate labeled data, this research adapts the pre-trained language model BERT-base-multilingual-cased, which was trained on a large multilingual corpus. The strategy involves two stages: first, the model is fine-tuned on a rich English-language spam dataset, and second, the trained model is then further fine-tuned using a much smaller Indonesian-language dataset. Quantitative evaluation results show that the model achieved very good and consistent performance in both languages. On the English dataset, the model reached an Accuracy of 0.9738 and an F1-score of 0.9436. More significantly, on the Indonesian dataset, the model achieved an Accuracy of 0.9492 with an F1-score of 0.9494. The comparable performance between the two languages, despite the Indonesian dataset being much smaller, proves that the semantic knowledge acquired from the source language (English) can be efficiently transferred for the same classification task in the target language (Indonesian). This research provides a strong demonstration of how transfer learning can bridge the data resource gap and has important implications for the development of NLP applications in the context of low-resource languages

Klasifikasi Buah dan Sayuran Multi-label Menggunakan CNN: Mengatasi Class Imbalance Dengan Focal Loss

Gita Ayu Syafarina — 2025-12-26

Investigates the effectiveness of Focal Loss as a solution to the problem of class imbalance in multi-label fruit and vegetable classification tasks. Using a ResNet50-based Convolutional Neural Network (CNN) architecture, two models were trained and evaluated: one using Focal Loss and another using Binary Cross-Entropy (BCE) Loss as a baseline. To address the availability of multi-label datasets, a synthetic multi-label dataset was created by combining images from existing single-label datasets. Experimental results show that the model trained with Focal Loss achieved an accuracy of 0.9390 and an F1-score of 0.9863, outperforming the BCE Loss model which only reached an accuracy of 0.8850 and an F1-score of 0.9718. The comparative analysis indicates that Focal Loss, with its ability to focus the training process on difficult examples, effectively addresses class imbalance and produces superior performance. This study concludes that Focal Loss is an effective tool for multi-label classification tasks and highlights the existing limitations, including the synthetic nature of the dataset and the limited training duration, which underscore the need for further research

Deteksi Serangan Dalam Ekosistem Iot Melalui Analisis Multi-Class Dengan Model Xgboost Dan Penerapan Teknik Imbalance Ratio Pada Dataset IoTID20

Januar Al Amien — 2025-12-28

This research focuses on attack detection in the Internet of Things (IoT) ecosystem using the XGBoost algorithm and the Imbalance Ratio technique on the IoTID20 dataset. The main goal is to overcome the problem of data imbalance that is common in IDS datasets and improve accuracy in classifying attack types. The methodology used includes data preprocessing, feature selection, and applying the Imbalance Ratio technique to handle class imbalance in the IoTID20 dataset. Next, the XGBoost model is implemented with the scale_pos_weight parameter to handle the class imbalance problem. This model is trained on training data and evaluated using metrics such as accuracy, precision, recall, and F1-score. The research results show that the combination of the XGBoost algorithm and the Imbalance Ratio technique is able to overcome data imbalance problems effectively. The resulting model achieved an accuracy rate of 99.32%, precision 99.32%, recall 99.32%, and F1-score 99.32% in classifying attack types on the IoTID20 dataset. These results demonstrate excellent capabilities in detecting attacks and distinguishing between normal and anomalous traffic in the IoT ecosystem. This research contributes to improving IoT network security by applying an effective Machine Learning approach to accurately detect attacks, while also addressing data imbalance problems that often occur in IDS datasets.

Implementasi logika fuzzy mamdani dan simple additive weighting (saw) pada sistem pakar berbasis web untuk deteksi dini gangguan neurologis

Muhammad Harits Firdaus Harits — 2025-12-28

Neurological disorders such as low back pain, vertigo, ischemic stroke, epilepsy, and peripheral neuropathy affect the central and peripheral nervous systems and have the potential to reduce quality of life and be fatal if not detected early. In Indonesia, the high prevalence is not balanced with access to early diagnosis due to limited medical personnel, costs, and waiting times. This study developed a web-based expert system for early detection of five neurological disorders using the Mamdani Fuzzy Method for inference and Simple Additive Weighting (SAW) for symptom ranking. The diagnosis process includes fuzzification, rule evaluation, aggregation, centroid defuzzification, and SAW calculation. The system was tested through black box testing and accuracy evaluation using MAE, RMSE, and F1 Score. The results showed an MAE value of 2.8%, RMSE 2.83%, and F1 Score 0.75, which proves the system is accurate, consistent with manual calculations, and easy to use. With a user-friendly interface, this system has the potential to be a pre-diagnosis tool that increases public awareness and supports medical personnel in decision-making.

Classification of Smartphone Addiction of Esa Unggul University Students Using Machine Learning and Sas-sv

Anggoro Verrel — 2025-12-29

The digital era has made smartphones an inseparable part of students' lives, but it also raises the risk of addiction that negatively impacts academic achievement and mental health. This research aims to develop and evaluate machine learning models capable of classifying the level of smartphone addiction among Esa Unggul University students. Data were collected from 32 student respondents through an online questionnaire adopting the validated psychometric instrument, the Smartphone Addiction Scale-Short Version (SAS-SV). Addiction levels were categorized into two classes: 'High', which refers to the gender-specific addiction risk threshold from Kwon et al. (2013), and 'Moderate', which includes scores below that threshold. Four classification algorithms—Logistic Regression, K-Nearest Neighbors (KNN), Decision Tree, and Random Forest—were implemented to compare their performance. To address class imbalance in the data, the SMOTE oversampling technique was applied to the training data. Model evaluation was based on accuracy, precision, recall, and F1-score. The research results show that the Logistic Regression model achieved the best performance with an accuracy of 1.0000 and an average F1-score of 1.00 on the test data. Nevertheless, it should be noted that this perfect performance was obtained from a very limited test data size (8 samples), so generalization requires further validation. Feature importance analysis from the Random Forest model identified that the question related to Planned tasks/work often interrupted by smartphone use (Q0) was the most dominant predictor. These results indicate that machine learning models based on psychometric scales have initial potential as a screening and exploratory tool to identify students at risk of smartphone addiction, but require extensive development and validation on larger data before practical implementation.

Studi Literatur: Perencanaan Arsitektur Fisik Server Big Data Manajemen Global Tapak Suci Putera Muhammadiyah

Rahmad Al Rian — 2025-05-29

The Indonesian Martial Arts School Tapak Suci Putera Muhammadiyah has been established in 22 countries. In its cadre formation process, Tapak Suci has 15 levels ranging from elementary students to great warriors, where each level will be achieved through a level promotion exam process. In terms of membership age, Tapak Suci has the youngest members at the elementary school level. In the process of building achievements, Tapak Suci regularly holds regional championship events organized by regional leaders, provincial levels organized by regional leaders, national and international levels organized by central leaders. In addition, Tapak Suci members also actively participate in pencak silat events organized by IPSI, Persilat and other pencak silat events. In the process of organizational management, Tapak Suci has a training branch management, regional leaders, regional leaders, regional representatives (overseas) and central leaders. This article is the result of a literature study to determine the server needed by Tapak Suci to carry out global management of its big data so that the development of Tapak Suci's human resources can be monitored in real-time.

Aplikasi Manajemen Keuangan Pribadi Berbasis Mobile Menggunakan Framework Flutter

Tubagus Muchamad Isnaeni — 2025-12-31

Personal financial management is a crucial aspect of maintaining individual financial stability. However, many people struggle to effectively record, manage, and monitor their income and expenses. This research aims to develop a mobile-based financial management application using the Flutter Framework, focusing on simple record-keeping (CRUD) features. It also features transaction categories, monthly summaries, and the user's final balance. This application is designed to help users easily record income and expenses, display a list of transactions, and allow for editing and deletion of data. With a minimalist approach, this application does not require an online connection, allowing it to be used offline and ensuring user data privacy. The database system used is SQLite, which is lightweight and suitable for mobile applications. The development method used is the Waterfall method, which includes requirements analysis, system design, implementation, testing, and maintenance. Test results show that this application is able to meet basic financial record-keeping needs with a simple and responsive interface. This application is expected to provide a practical solution for individuals seeking a simple financial management tool and serve as a foundation for the development of more complex features in the future.

Perencanaan SI pada Website Alat Kopi Gaharu Menggunakan Analisis SWOT dan CSF

Muhammad Rafli — 2025-12-31

The utilization of information systems and information technology (IS/IT) by MSMEs is still not optimal, especially in supporting website-based marketing and sales. UMKM Gaharu Coffee Tools faces problems such as limited website features, less-than-optimal access speed, and the lack of integration between the ordering and payment systems, which impacts operational effectiveness and business competitiveness. This research aims to design an SI/IT strategic plan for the Gaharu Coffee Tools website to enhance business competitiveness. The research methods used include interviews and observations, which were analyzed using the SWAT and Critical Success Factors (CSF) methods. The results of the SWAT analysis show a difference of 1.38 between strengths and weaknesses and a difference of 1.10 between opportunities and threats, placing the website in the aggressive (SO) quadrant. The CSF analysis yielded three main priorities: capitalizing on the trend of home brewing, expanding the online market internationally, and managing a structured product catalog. In conclusion, targeted website development aligned with SI/IT priorities has the potential to enhance user experience, expand market reach, and sustainably strengthen the competitiveness of MSMEs.

Implementasi CNN untuk Identifikasi Penyakit Daun Cabai

Aqmal Salya Nur Alamsyah — 2025-12-31

Disease detection in chili plants is a crucial step in preventing damage that can reduce productivity and cause economic losses for farmers. This study presents the design of an Android-based application called Chili Leaf Disease App that can automatically detect chili leaf diseases. The application uses a Convolutional Neural Network (CNN) algorithm with the MobileNetV2 architecture to classify leaf diseases through images captured directly from the camera or uploaded from the gallery. The dataset used consists of 4,000 chili leaf images across four disease classes. Testing results show that the model achieves an accuracy of 97.5%. The system was developed using the Rapid Application Development (RAD) method, chosen for its shorter development cycle, flexibility, and ability to increase user involvement. This approach enables efficient, fast, and user-responsive application development. The application is expected to help farmers detect diseases early and take preventive action more quickly to maintain plant health.

Analysis Application of the Random Forest Algorithm in Weather Forecast Classification

Deny Saputra Saputra — 2025-12-31

Weather plays an important role in various aspects of life, such as agriculture and transportation. However, weather prediction remains challenging because it is influenced by many complex factors. Extreme weather events, such as storms and floods, can cause significant losses, making accurate weather forecast classification systems essential. This study applies the Random Forest algorithm to improve prediction accuracy and optimizes it using Grid Search Cross Validation. The method used is CRISP-DM, consisting of six main stages. The data were obtained from the Meteorological, Climatological, and Geophysical Agency (BMKG), containing features such as temperature, humidity, wind speed, cloud cover, visibility, and wind direction, with the labels Weather Condition and Region Name serving as indicators of the classified weather category and location. The final evaluation uses a confusion matrix, yielding an accuracy of 98.84% on the training data and 95.33% on the testing data, demonstrating stable performance and strong generalization capability.

Implementation of Convolutional Neural Network for Disease Detection in Mobile-Based Clove Leaves

Satria Junmulyana — 2025-12-31

Clove (Syzygium aromaticum) is a spice crop that has high economic value, but faces serious threats from various diseases that can reduce yields. Early detection of disease in clove plants is very important to prevent greater losses. This research aims to develop a disease detection system for clove plants using Convolutional Neural Network (CNN) implemented in a mobile application. This method is expected to provide a faster and more accurate solution compared to traditional detection methods that are often inefficient. This research was conducted by collecting datasets of infected and healthy clove leaf images, which were then used to train the CNN model. The results show that the developed CNN model is able to achieve high disease detection accuracy, and can be integrated with mobile technology to facilitate farmers in identifying diseases in real-time. Thus, this research not only contributes to increasing agricultural productivity, but also supports the application of digital technology in the agricultural sector. The results of this research are expected to benefit farmers, researchers, and the agricultural industry as a whole.