Pseudocode

Prototype Sistem Monitoring Status Pintu Berbasis IoT Dengan Notifikasi Web dan Telegram

Apriansa Arwandi Panjaitan — 2026-02-28

In the era of the Industrial Revolution 4.0, the Internet of Things (IoT) has become an innovative solution for enhancing home security. This study designs and implements an IoT-based monitoring door status system using the ESP8266 microcontroller combined with a magnetic sensor. The system enables real-time monitoring and notifications to users via Telegram and provides door status display through a website. With an integrated alarm feature to prevent unauthorized access, this system is expected to enhance home security at a more affordable cost compared to conventional security systems. Testing results indicate that the system accurately detects door status changes with a notification response time of less than 2 seconds. These findings demonstrate the system’s effectiveness in improving security and user conveniencel.

Klasifikasi Severity Level Diabetic Macular Edema Berbasis ResNet-50

I Made Dendi Maysanjaya — 2026-02-28

Diabetes is one of the most common diseases people suffer from today, and it can lead to complications such as blindness, heart disease, and kidney failure. The condition of blindness caused by this disease is known as diabetic retinopathy (DR). An ophthalmologist will use a fundus camera to examine the retina, looking for several clinical features, such as microaneurysms (MA), hemorrhages (HM), cotton-wool spots (CWS), and exudates. Based on these clinical symptoms, clinicians then determined the patient's level of diabetic macular edema (DME) severity. Although several studies have applied CNN-based architectures for diabetic retinopathy detection, limited attention has been given to the impact of dataset imbalance handling on DME severity classification, particularly using ResNet-50. This study highlights the significant impact of extensive data augmentation on classification performance in imbalanced DME datasets. Evaluate performance using the accuracy, precision, and recall metrics. We used the IDRiD dataset, which consists of 516 images split into a training set of 413 and a test set of 103. IDRiD divides the dataset into three classes, namely normal, moderate DME, and severe DME. In the preprocessing stage, we enhanced contrast using CLAHE and resized the images to 224x224 pixels. To address the imbalance, we applied 11 data augmentation methods. We experimented by comparing the performance of two models: one with and one without dataset augmentation. Based on the test results, the best performance was obtained with the model that included dataset augmentation, achieving an accuracy of 0.5961, a precision of 0.63, and a recall of 0.61, while the baseline model (without dataset augmentation) gained 0.4553, 0.36, and 0.34 for the accuracy, precision, and recall, respectively.

Pengembangan Model Pengenalan Huruf SIBI Pada Kondisi Low-Light Berbasis Convolutional Neural Network

Francisco Francisco — 2026-02-28

Deaf and speech-impaired individuals in Indonesia face communication barriers due to limited public understanding of sign language. In real use, SIBI communication often occurs in dim lighting, yet recognition models are mainly evaluated under normal illumination, motivating robust low-light recognition. This study develops a CNN model based on MobileNetV2 to recognize SIBI (Indonesian Sign Language System) letter gestures under low-light conditions (50-100 lux). The dataset comprises 5,579 images of 26 SIBI letters, divided stratified 80:10:10. The methodology includes preprocessing with Bilateral Filter, CLAHE in LAB color space, and Adaptive Gamma Correction, plus transfer learning and fine-tuning with data augmentation. Evaluation results show 97.13% test accuracy, with most errors among similar letters. Real- time testing is stable within 50-100 lux, though accuracy decreases below 50 lux or with shadows. These findings indicate that the proposed preprocessing methods and MobileNetV2 CNN maintain reliable SIBI recognition in low-light environments.

Analisis Sentimen Ulasan Aplikasi Access by KAI Menggunakan Algoritma Naïve Bayes

Beni Ariansyah — 2026-02-28

The Access By KAI application, developed by PT Kereta Api Indonesia (Persero), allows users to purchase train tickets via mobile devices. This study aims to perform sentiment analysis on user reviews of the Access By KAI application using the naive Bayes algorithm. Data processing was carried out through stages such as case folding, cleaning, tokenizing, stopword removal, and stemming, and evaluation using metrics of accuracy, precision, recall, and F1-score showed that the naive Bayes algorithm provides satisfactory results. The study results indicate that the naive Bayes algorithm is able to classify reviews with an accuracy rate of up to 68% with a precision of 83% for the positive class, 59% for the negative class, and 79% for the neutral class; recall of 67% for the positive class, 93% for the negative class, and 42% for the neutral class. From these results, it is expected to help developers identify the aspects most complained about by users and improve service quality.

Penerapan Algoritma Random Forest dan Support Vector Machine Untuk Deteksi Distributed Denial of Service

Asep Ririh Riswaya — 2026-02-28

Distributed Denial of Service (DDoS) attacks significantly threaten the availability of modern network services by overwhelming resources with malicious traffic. The increasing complexity of these attacks, including multi-vector and low-rate variants, reduces the effectiveness of traditional detection methods based on signatures and static rules. This study explores the effectiveness of Random Forest (RF) and Support Vector Machine (SVM) algorithms in detecting DDoS attacks using the CICDDoS2019 dataset, focusing on the impact of various decision threshold values on performance. The CICDDoS2019 dataset consists of 431,371 network traffic flows with 80 numerical features. Preprocessing involves eliminating null values, standardizing numerical attributes, and encoding labels into binary classifications of normal and DDoS traffic. The dataset is then divided into training and testing sets at a 70:30 ratio. Performance evaluation is done using a confusion matrix to calculate accuracy, precision, recall, and F1-score. Results show that both algorithms perform well, but Random Forest offers greater consistency, with a threshold of 0.5 achieving the best balance in metrics.

Analisis Penerapan Aplikasi Media Pembelajaran Interaktif model 3D Berbasis Augmented Reality

Jimmy Pratama — 2026-02-28

Technological developments have encouraged the use of Augmented Reality (AR) as an interactive learning medium in education. However, learning 3D animation using Autodesk Maya is still considered difficult by beginner students due to the complexity of the features and interface. This study aims to develop and analyze the benefits of an interactive learning media application based on markerless Augmented Reality (AR) in recognizing icons and functions of the Autodesk Maya Toolbox. The research method used is applied research with a qualitative approach and the Multimedia Development Life Cycle (MDLC) development method. The application was developed using Unity, Vuforia, and Autodesk Maya, and tested on 30 students through interviews. The results of the study show that the markerless AR application is able to help students understand and recognize the Autodesk Maya toolbox more easily through interactive and clear 3D object visualization. Thus, this application is considered useful as a supporting medium for innovative and interactive 3D animation learning.

Implementasi Sistem Verifikasi dan Pengenalan Wajah Berbasis YOLOv8n dan VGG-Face untuk Presensi Perkuliahan

Cahyaningsih Listya Suhardi — 2026-02-28

This research presents the development of a facial recognition and verification system that aims to address attendance record falsification in university lectures, a persistent challenge in non-biometric attendance management. The proposed framework integrates the efficiency of YOLOv8n for facial detection with the strong feature representation capability of VGG-Face. The system applies image augmentation to create varied facial embeddings, uses Cosine Similarity for identity verification, and evaluates its performance through accuracy, precision, recall, and F1-Score metrics. Experimental evaluations on student facial datasets captured under different lighting conditions, poses, and viewing angles show that the system achieves a stable accuracy of around 90 % without augmentation, increasing to 97 % with augmentation, which enhances overall stability and reliability. These results demonstrate that the integration of YOLOv8n and VGG-Face offers an effective and dependable solution for strengthening the security and credibility of facial recognition-based attendance systems in academic settings.

Klasifikasi Burnout Mahasiswa menggunakan Algoritma Random Forest dengan Analisa Feature Importance

Andrean yuriko — 2026-02-28

Prolonged stress in higher education environments can lead to emotional, physical, and mental exhaustion among students, negatively affecting concentration, academic performance, and mental health. This study aims to classify student burnout levels using the Random Forest algorithm and to analyze model performance as well as dominant contributing factors through a feature importance approach. The dataset was obtained from the Kaggle platform and consisted of 1,100 samples covering psychological, physiological, environmental, academic, and social variables. The research methodology included data preprocessing, data balancing using the Synthetic Minority Over-sampling Technique (SMOTE), and data splitting with six scenarios: 70:30, 75:25, 78:22, 80:20, 83:17, and 87:13. Experimental results showed that all splitting scenarios produced relatively stable performance, with the 80:20 split achieving the highest accuracy of 89%, while other scenarios ranged from 85% to 87%. Feature importance analysis indicated that physiological factors, particularly blood_pressure, had the most significant contribution to student burnout classification.

Analisis Komparatif Efektivitas Keamanan dan Penerimaan Pengguna Pada Protokol Presensi Digital Berbasis QR-Code Statis dan Dinamis

Dwi Purnomo Putro — 2026-02-28

QR code–based digital attendance systems are widely used in higher education, yet they continue to face security challenges such as attendance fraud and inaccurate time recording. This study aims to comparatively evaluate the effectiveness of static and dynamic QR code–based attendance protocols by examining system security and user acceptance. A quantitative approach was employed through a field experiment and a survey grounded in the Technology Acceptance Model (TAM), encompassing Perceived Usefulness, Perceived Ease of Use, and Behavioral Intention. Data were analyzed using independent samples t-tests and Cohen’s d effect size measures. The results show that the dynamic QR code protocol significantly outperforms the static QR code across all TAM constructs (p < 0.001), with very large effect sizes (d > 1.0). These findings indicate that enhanced security mechanisms do not reduce usability, but instead strengthen user acceptance and adoption intention.

Analisis Prediksi Probabilitas Otomatisasi Pekerjaan Tahun 2030 Menggunakan Algoritma Linear Regression Dan Gradient Boosting

Muhammad Arif Billah Natagama — 2026-02-28

The rapid development of artificial intelligence and automation is expected to significantly impact future employment. This study aims to predict job automation probability in 2030 using supervised learning methods. A public dataset containing job types, education levels, and automation probabilities was utilized. Linear Regression and XGBoost Regressor were employed to build and compare predictive models. The research process included data preprocessing, training–testing data split, model training, and performance evaluation using Root Mean Square Error (RMSE) and coefficient of determination (R²). Experimental results indicate that XGBoost outperforms Linear Regression by achieving lower RMSE and higher R² values. This study provides insights into automation risks and may support workforce skill development planning.

Sistem Informasi Geografis Berbasis Web Untuk Pemetaan Penerima BPJS Dengan Algoritma Dijkstra

Muhamad Reynaldi Rendi — 2026-02-28

The Sungai Raya Village apparatus faces challenges in managing BPJS beneficiaries due to the lack of mapping and route optimization tools. This research aims to build a web-based Geographic Information System (GIS) to visualize recipient locations and implement Dijkstra's Algorithm for determining the shortest visit routes. Developed using Laravel, MySQL, and Leaflet.js following the Waterfall methodology, the system successfully maps data with hamlet-based filtering features. Testing confirms that Dijkstra's Algorithm generates valid routes, which were verified against Google Maps. Black Box testing showed all functions perform correctly, and User Acceptance Testing (UAT) yielded a score of 81.75%. These results indicate the system effectively supports the efficiency of data verification and field visit planning.

Implementasi Metode Topsis Pada Sistem Pendukung Keputusan Penentuan Prioritas Penerima Bantuan Mesin Untuk Sentra IKM Berbasis Website

Exca Wella Monica — 2026-02-28

The industrial sector is one of the economic sectors being developed in Indonesia as a driving force for the progress of other economic sectors. One form of this is the Small and Medium Industries (SMEs), which are quite dependent on government assistance, both in terms of capital and equipment. One of the government programs to advance SMEs is the provision of machinery assistance to SME centers. In practice, the selection process for machine assistance recipients at the Bengkulu Industry and Trade Office still uses Microsoft Excel, which has limitations, including the risk of human error, logical errors in formulas, differences in results due to rounding numbers, and a lack of objectivity in assessments. Therefore, a Decision Support System (DSS) is needed to minimize errors, producing more objective, transparent, and accurate decisions. This study uses the TOPSIS method to calculate the preference value of SMEs and data aggregation techniques (averages) to calculate the final value of SME centers. The resulting output is a ranking of SME centers eligible for machine assistance. Of the total of 26 comparisons between the rankings from the system and the rankings from the Industry and Trade Office, 20 rankings were obtained that were the same, so that the system accuracy reached 76.92%. Furthermore, the system's feasibility test achieved an interval score of 4.6, or 92%, which is considered very good. The final result of this research is the creation of a website-based decision support system for prioritizing recipients of machine assistance for small and medium enterprises SME centers.

Perancangan Sistem Informasi Geografis (SIG) untuk Pendataan Jaringan Kabel pada Area Distribusi

Aldi Huseini Milyanto — 2026-02-28

This study focuses on the development of a Geographic Information System (GIS) designed to support the recording of cable network data in distribution areas. One of the main challenges encountered by network managers is the lack of an effective and accurate method for mapping the locations of network infrastructure. To address this issue, a web-based GIS is proposed to enable more structured processes of data recording, monitoring, and maintenance of poles and cable networks. The system development adopts a prototype-based approach, which includes requirement analysis, system modeling, implementation, and testing phases. The application is implemented using HTML, PHP, MySQL, and Leaflet.js as a digital mapping library. Spatial data are collected through field surveys and converted into geographic coordinates for visualization on an interactive map. The developed system is able to display pole locations and cable routes in real time, thereby supporting asset management and distribution network analysis. Overall, the system provides an effective solution for managing cable network infrastructure data.