Journal of Digital Ecosystem for Natural Sustainability

Arsitektur Hybrid Vision Transformer–ConvNeXt dengan Multi-Task Focal Loss dan Medical Test-Time Augmentation untuk Klasifikasi Lesi Kulit Berbasis Citra

2025-12-27T19:43:05+00:00

Dermatoscopy image-based skin lesion classification is a challenge in dermatology due to the high visual variation between lesion types and the imbalanced class distribution in the dataset. In this study, a Hybrid Vision Transformer–ConvNeXt architecture is proposed, combining the global attention capability of Vision Transformer (ViT) and the spatial feature representation of ConvNeXt, to improve the classification performance of skin lesion images on the HAM10000 dataset. This study also applies Multi-Task Focal Loss, auxiliary classifier, and Weighted Random Sampler to effectively address the class imbalance. In addition, the Medical Test-Time Augmentation (TTA) approach is used in the inference stage to improve the stability of predictions. The model is trained using a two-stage strategy (head training and full fine-tuning), as well as optimization based on AdamW and Cosine Annealing Warm Restarts. The test results show that the proposed model successfully achieves a validation F1-Score of 0.8723, and after TTA it increases to 0.90, surpassing the baseline of ViT and single ConvNeXt. These findings indicate that the integration of ViT–ConvNeXt with loss strategy and medical TTA is able to significantly improve the performance of skin lesion classification, and has the potential to be applied as a clinical diagnosis support system.

Perancangan User Interface dan User Experience Pada Sistem Inventaris Menggunakan User Centered Design

2025-12-16T23:47:23+00:00

With the growing influence of digital technology, the demand for a more efficient and accurate inventory system has become increasingly relevant, especially for companies that still rely on semi-manual record-keeping. This study was conducted at PT Kanindo Mandiri Sukses, where the current inventory process is still handled conventionally. Such a system often leads to difficulties in tracking data and increases the chance of input errors. To resolve these issues, this research designed a web-based inventory system tailored to user needs by applying the User Centered Design (UCD) approach. The design process involved identifying users, formulating requirements, creating a prototype interface using Figma, and evaluating the user experience through the User Experience Questionnaire (UEQ) method. The evaluation results revealed high scores in five UX aspects: Attractiveness, Perspicuity, Efficiency, Dependability, and Stimulation. The Novelty aspect scored moderately, indicating opportunities for improvement in innovation. Feedback from four internal employees indicated that the system helped simplify data entry and improved the efficiency of daily operations. It is expected that this system will support a more structured and effective inventory process in the company.

Glambot SNAFH-R1: Lengan Kamera 4-DOF Biaya Rendah untuk Photobooth Indoor Berbasis Pengolahan Citra

2025-12-19T07:54:26+00:00

This paper presents the design and implementation of SNAFH-R1, a low-cost four-degree-of-freedom (4-DOF) glambot camera arm intended for indoor photobooth applications. The system integrates a Python-based image processing module on a laptop with an Arduino Uno R3 controller that drives three high-torque RDS3225MG servos and one MG995 servo via a PCA9685 PWM driver. A USB camera (Logitech C270) is used to capture video frames that are processed in real time using grayscale conversion, illumination enhancement, noise reduction, and human detection based on a cascade classifier. The centroid of the detected region of interest is then converted into position offsets and mapped to target joint angles, which are applied to the servos using clamped and interpolated commands to produce smooth camera motion. A BH1750 light sensor and relay-controlled LED lamp are employed to automatically stabilize local illumination around the subject. Experimental observations in an indoor photobooth scenario show that SNAFH-R1 can maintain the subject’s head and upper body within the camera frame at short to medium distances, while producing reasonably smooth camera trajectories and operating stably during short test sessions. These results indicate that the proposed system is a feasible low-cost alternative for automated glambot-style camera movements in small to medium-scale events and as a teaching platform for robotics and computer vision.

Prediksi Kelompok Usia Pengguna Netflix Menggunakan Metode Random Forest Berdasarkan Analisis Genre Tontonan dan Perilaku Pengguna

2025-12-27T19:44:05+00:00

The accuracy of user demographics, particularly age, on video streaming platforms is often compromised by the widespread practice of shared accounts. This study addresses this challenge by implicitly classifying user age groups (Youth, Young Adult, Adult, Middle-Aged, Senior) based solely on behavioral data, including viewing genre frequency, sentiment analysis of reviews, and expenditure patterns. The core methodology employs a Random Forest Classifier optimized with SMOTE (Synthetic Minority Over-sampling Technique) to mitigate the severe class imbalance present in the dataset. The initial Baseline Model performed poorly, achieving only 40,13% accuracy and failing to identify minority classes. After implementing SMOTE and hyperparameter tuning, the Final Model demonstrated significant improvement, achieving an Accuracy of 79,26%. The engineered feature, Spend per Person, was identified as the most dominant predictor, validating the approach of using economic factors to differentiate genuine individual usage. Crucially, the model showed exceptional reliability in detecting sensitive age segments, such as Youth (F1-Score 0,88) and Seniors (F1-Score 0,75). This research provides an effective data-driven solution for enhancing age-based content personalization and parental control features.

Perancangan dan Evaluasi NUMORi sebagai Media Pembelajaran Numerasi Montessori Berbasis Cetak 3D untuk Anak Usia Dini

2025-12-29T17:27:23+00:00

The long-term objective of this research is to develop an applied scientific design model in the field of multimedia engineering for producing 3D-based educational products aligned with the Montessori learning philosophy. This study aims to contribute to the establishment of a standardized design framework for early childhood numeracy learning media that integrates scientific rigor, aesthetic value, and functional effectiveness. The specific target of this research is the development of a physical prototype called NUMORi (Numerical Montessori 3D Model), an interactive learning aid produced using 3D printing technology and designed through a scientific design approach, along with the acquisition of intellectual property rights (IPR) for the educational product and the development of an implementation guidebook for kindergarten learning environments. The research employs a scientific design methodology consisting of needs observation, design problem formulation, 3D model design hypothesis development, experimentation through modeling and 3D printing, and evaluation through trials in Montessori schools in Batam City. Data are collected through observations, teacher interviews, and expert assessments. The findings are expected to provide both theoretical and practical contributions to the development of scientific design methods in interactive multimedia and education.

Pengembangan Chatbot Pelayanan Akademik Berbasis Arsitektur Hybrid Laravel dan Flask dengan Pattern Matching Regular Expression

2025-12-19T09:31:59+00:00

Academic information services demand high precision and near-instant responsiveness. While chatbot adoption in higher education is growing, existing approaches face a technical dilemma: deep learning-based Natural Language Processing (NLP) models often require significant computational resources and are prone to "hallucinations" (inaccurate answers), whereas traditional rule-based systems are rigid and difficult to maintain. This study addresses these limitations by proposing a Smart Chatbot utilizing a Hybrid Architecture that decouples the user interface (Laravel-based) from the processing logic (Python Flask-based) via REST API. The core intelligence relies on an Optimized Hierarchical Regular Expression (Regex) method Deterministicigned to provide deterministic and accurate responses for administrative queries without the computational overhead of complex AI models. Performance testing using Black Box methodology on 50 query samples demonstrated a 92% accuracy rate with an exceptionally low average latency of 0.7 seconds. These results confirm that a hybrid approach combined with deterministic pattern matching offers a superior balance of speed, accuracy, and resource efficiency for academic environments compared to resource-intensive NLP models

Penerapan Combined Compromise Solution dalam Pendukung Keputusan Program Bantuan Pelaku UMKM Kota Lubuklinggau

2025-12-30T14:45:13+00:00

Micro, Small, and Medium Enterprises (MSMEs) play a significant role in strengthening regional economies and providing employment opportunities, particularly in developing cities. The Government of Lubuklinggau City implements assistance programs for MSME actors aimed at enhancing business capacity, competitiveness, and sustainability through funding support, training, and other supporting facilities. In practice, the selection process for aid recipients is often challenging due to multiple evaluation criteria, diverse data, and the potential for subjectivity in assessments. To address these challenges, this study applies the Combined Compromise Solution (COCOSO) method as an optimization approach in the decision-making process. COCOSO is a metaheuristic algorithm that integrates parallel cooperation and information exchange mechanisms among subpopulations, producing more optimal solutions and accelerating convergence. In this research, COCOSO is utilized to process various selection criteria,. The list of alternatives was obtained from the Department of Cooperatives, Small and Medium Enterprises of Lubuklinggau City. The experimental results indicate that this approach provides fairer, more objective, and transparent decision-making compared to traditional methods. Therefore, COCOSO can serve as an effective decision support system to assist the government in distributing MSME assistance programs more accurately and precisely in Lubuklinggau City.

Simulasi Rumah Pintar Berbasis IOT Menggunakan Aplikasi Cisco Paket Tracer

2025-12-29T17:10:59+00:00

The development of Internet of Things (IoT) technology has driven the creation of smart home systems, which enhance comfort, security, and energy efficiency for occupants. This study aims to simulate an IoT-based smart home system using Cisco Packet Tracer as a visualization and network testing tool. In this simulation, various smart devices—such as automated lights, temperature sensors, fans, and digital doors—were configured and integrated using IoT communication protocols. The simulation results demonstrate that Cisco Packet Tracer can be effectively used to model interactions between devices in a smart home network and test automated responses based on input from installed sensors. This simulation is expected to serve as a foundation for developing more complex smart home systems and as a learning medium for understanding basic IoT concepts in home automation