Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi)
Not a member yet
1071 research outputs found
Sort by
Modeling and Deploying RESTful Services with SOMF-Based SOA: A Case Study in the Credit Guarantee Industry
No full textThe integration of electronic systems across financial institutions poses significant challenges, particularly when legacy architectures rely on siloed, point-to-point connections. This often leads to what is commonly known as "spaghetti integration," where changes in one system can trigger unintended disruptions in others. This study addresses such integration issues within the Kredit Usaha Rakyat (KUR) credit guarantee service of an Indonesian credit guarantee institution by implementing a Service-Oriented Architecture (SOA) approach, guided by the Service-Oriented Modeling Framework (SOMF). This study aims to improve system performance, scalability, and regulatory adaptability through a structured, multi-phase methodology based on SOMF: conceptualization, discovery and analysis, business integration, logical design, and logical architecture. Data for the study were drawn from system documentation, national regulatory requirements (e.g., Coordinating Minister Regulation No. 1/2023), and the evaluation of service interactions via RESTful APIs using lightweight JSON formatting. These findings demonstrate that the adoption of SOA with SOMF enables the development of modular, interoperable, and adaptable services. This approach reduces redundant processes, enhances real-time data flow, and strengthens integration between the guarantee institution and its partner banks. The resulting system aligns with modern digital governance requirements and provides a sustainable foundation for future growth and compliance
Development of IoT-based Automatic Water Drainage System on Fishing Boat to Improve Operational Efficiency
Full text linkThe profession of fishermen requires a reliable system to remove stagnant water from fishing boats, as manual drainage is time-consuming and inefficient. This study proposes an IoT-based automatic water drainage system without using an inverter or ultrasonic sensor, offering a cost-effective alternative. The system utilizes a water level sensor and a DC water pump, controlled via a smartphone application. The research model used is the Research and Development (R&D) model, through several stages, namely potential and problems, initial data needs, prototype creation, prototype validation, prototype revision, validation, implementation. Problems occur at the prototype stage, problems that must be revised include aspects of wiring, Power Suitability, Water Level Sensor Test, and the configuration of the relay used. The IOT-based automatic water drainage system can function based on the results of white-box testing including Hardware Implementation, Software Implementation, Implementation of Application Usage, and Automatic Drainage System Testing. This is indicated by the results of the Liquid Water Level Sensor Functionality test, DC Water Pump Functionality Test, Solar Panel and Battery Functionality Test, and IOT Functionality Test. IOT-based automatic water discharge systems on fishing boats are more efficient and cost-effective in the long run, although diesel engines offer more reliability under adverse weather conditions or in places with limited access to sunlight
Optimizing Sentiment Analysis for Lombok Tourism Using SMOTE and Chi-Square with Machine Learning
Full text linkTourism is a vital economic sector for Lombok Island, which is renowned for its natural beauty and cultural richness as a top destination. The rapid growth of tourism in Lombok requires a deep understanding of tourists' perceptions and sentiments to ensure an optimal service quality. The sentiment analysis of online reviews is valuable for identifying service strengths and weaknesses and addressing tourists' needs more effectively. This not only enhances tourist satisfaction, but also aids in the design of more effective marketing strategies. However, text data analysis from online reviews presents unique challenges such as noise, class imbalance, and numerous features that may affect classification results. Therefore, this study aims to classify tourist sentiment toward Lombok tourism using machine learning methods combined with feature selection and oversampling techniques. This study focuses on optimizing sentiment analysis of tourism-related tweets using a combination of SMOTE oversampling and Chi-Square feature selection on improving classification performance without hyperparameter tuning. The study applies machine learning methods, such as SVM and Naïve Bayes, with feature selection and oversampling using Chi-Square and SMOTE. The dataset used was sentiment data regarding Lombok tourism obtained from Twitter in 2023, consisting of 940 instances divided into three classes: Negative, Neutral, and Positive. The research findings show that the use of SMOTE and Chi-Square can improve the accuracy of the SVM and Naive Bayes methods. Without optimization, the SVM method achieved an accuracy of 73.93% and a Naive Bayes of 67.02%. After optimization with SMOTE and Chi-Square, the accuracy increased for SVM by 90% and Naive Bayes by 84% to classify tourist sentiment towards Lombok tourism. The implications indicate that combining data balancing using SMOTE with feature selection via Chi-Square effectively improves the performance of sentiment classification models for tourist opinions on Lombok's tourism
Development of MongoDB-based Gait System with Interactive Visualization for Clinical Analysis
Full text linkGait analysis is a crucial aspect of biomechanics and medical rehabilitation, used to detect movement disorders, assess therapy effectiveness, and understand human walking patterns. In Indonesia, gait research remains limited, with most data sourced from abroad, which may not reflect the characteristics of the local population. This study uses data from Vicon camera recordings that track marker movements on the subject's body and convert them into kinematic data in spatial coordinates, stored in Excel files. To support clinical applications, an efficient system is needed to manage gait data and present analysis results interactively. Therefore, a MongoDB-based gait data management system was developed due to its flexibility in handling unstructured data and scalability. The system was designed to preprocess gait data and display the results through an interactive Streamlit dashboard. The analysis involved calculating gait angle parameters, visualized in a gait cycle angle graph and analyzed statistically using mean and standard error to improve interpretation accuracy. Testing shows that the system can store data in an average of 1.52 seconds, retrieve it in 3.598 seconds, and render visualizations in 0.192 seconds, with high accuracy and only a 0.1-degree error between the input and output. This system effectively addresses the challenge of managing local gait data and supports comprehensive biomechanical analysis, enabling clinicians to make informed decisions regarding rehabilitation needs based on deviations from normal gait angle ranges
Benchmarking Metaheuristic Algorithms Against Optimization Techniques for Transportation Problem in Supply Chain Management
Full text linkThe optimization of transportation problems plays a significant role in supply chain management (SCM), where minimizing costs and improving efficiency are mandatory. The transition from manual methods to advanced computational approaches, such as metaheuristic algorithms, enhances decision-making and consolidates operations within SCM. Malaysia's transportation system has been confronting crucial challenges, characterized by congested roadways, limited rail connectivity and inefficient port operations, which interfere with the fluidity of goods and supply chain efficiency. This highlights the critical need for optimization techniques to enhance competitiveness and efficiency in the evolving SCM landscape. The research aims to explore the application of metaheuristic algorithms, with the Modified Distribution (MODI) method as the benchmark while employing the NorthWest Corner Method (NWCM) to obtain an initial feasible solution, to evaluate their performance in optimizing transportation problems. Metaheuristic algorithms, specifically Simulated Annealing (SA) and Particle Swarm Optimization (PSO), are implemented to explore alternative near-optimal solutions and assess the performance in terms of cost accuracy and computational efficiency. The results indicate that SA achieves a deviation of 12.92% in cost accuracy compared to the optimal MODI method, making it suitable for scenarios where precision is critical, whereas PSO which is 296.92 seconds faster, is ideal for time-sensitive applications. Finally, this study encourages future studies to explore additional algorithms, external factors and broader applications for enhanced real-world relevance and scalability to accentuate the potential of metaheuristic algorithms
Comparison of Transfer Learning Architecture Performance for Indonesian Auction Object Classification
No full textThe Indonesian auction, one of the sources of Indonesia's income for Non-Tax State Revenue (PNBP), faces challenges in accurately classifying auction objects, limiting revenue optimisation. This research aims to compare the performance of several transfer learning architectures on the Indonesian Auction Object Dataset, which includes categories such as Buildings, Cars, Motorbikes, and Salvage Materials. Seven pre-trained transfer learning models—MobileNetV2, NASNetMobile, EfficientNetV2B0, DenseNet121, Xception, InceptionV3, and ResNet50V2—were evaluated against a baseline model, focusing on validation accuracy, model size, and computational efficiency. MobileNetV2, NASNetMobile, DenseNet121, Xception, InceptionV3, and ResNet50V2 all achieved 100% validation accuracy, outperforming the baseline model's 96.5% accuracy. MobileNetV2 stands out for its efficiency, reaching 100% accuracy in just eight epochs with a compact model size of 11.1 MB. In contrast, EfficientNetV2B0 performed poorly on this dataset, achieving only 25% validation accuracy. These findings confirm that transfer learning architectures can significantly improve auction object classification accuracy while reducing the model size and training time, highlighting the benefit of transfer learning for optimising Indonesian auction systems
Evaluating Transformer Models for Social Media Text-Based Personality Profiling
No full textThis research aims to evaluate the performance of various Transformer models in social media-based classification tasks, specifically focusing on applications in personality profiling. With the growing interest in leveraging social media as a data source for understanding individual personality traits, selecting an appropriate model becomes crucial for enhancing accuracy and efficiency in large-scale data processing. Accurate personality profiling can provide valuable insights for applications in psychology, marketing, and personalized recommendations. In this context, models such as BERT, RoBERTa, DistilBERT, TinyBERT, MobileBERT, and ALBERT are utilized in this study to understand their performance differences under varying configurations and dataset conditions, assessing their suitability for nuanced personality profiling tasks. The research methodology involves four experimental scenarios with a structured process that includes data acquisition, preprocessing, tokenization, model fine-tuning, and evaluation. In Scenarios 1 and 2, a full dataset of 9,920 data points was used with standard fine-tuning parameters for all models. In contrast, ALBERT in Scenario 2 was optimized using customized batch size, learning rate, and weight decay. Scenarios 3 and 4 used 30% of the total dataset, with additional adjustments for ALBERT to examine its performance under specific conditions. Each scenario is designed to test model robustness against variations in parameters and dataset size. The experimental results underscore the importance of tailoring fine-tuning parameters to optimize model performance, particularly for parameter-efficient models like ALBERT. ALBERT and MobileBERT demonstrated strong performance across conditions, excelling in scenarios requiring accuracy and efficiency. BERT proved to be a robust and reliable choice, maintaining high performance even with reduced data, while RoBERTa and DistilBERT may require further adjustments to adapt to data-limited conditions. Although efficient, TinyBERT may fall short on tasks demanding high accuracy due to its limited representational capacity. Selecting the right model requires balancing computational efficiency, task-specific requirements, and data complexity
Real-time Emotion Recognition Using the MobileNetV2 Architecture
Full text linkFacial recognition technology is now advancing quickly and is being used extensively in a number of industries, including banking, business, security systems, and human-computer interface. However, existing facial recognition models face significant challenges in real-time emotion classification, particularly in terms of computational efficiency and adaptability to varying environmental conditions such as lighting and occlusion. Addressing these challenges, this research proposes a lightweight, yet effective deep learning model based on MobileNetV2 to predict human facial emotions using a camera in real time. The model is trained on the FER-2013 dataset, which consists of seven emotion classes: anger, disgust, fear, joy, sadness, surprise, and neutral. The methodology includes deep learning-based feature extraction, convolutional neural networks (CNN), and optimization techniques to enhance real-time performance on resource-constrained devices. Experimental results demonstrate that the proposed model achieves a high accuracy of 94.23%, ensuring robust real-time emotion classification with a significantly reduced computational cost. Additionally, the model is validated using real-world camera data, confirming its effectiveness beyond static datasets and its applicability in practical real-time scenarios. The findings of this study contribute to advancing efficient emotion recognition systems, enabling their deployment in interactive AI applications, mental health monitoring, and smart environments. Real-world camera data is also used to evaluate the model, demonstrating its usefulness in real-time applications and its efficacy beyond static datasets. The results of this work advance effective emotion identification systems, making it possible to use them in smart settings, interactive AI applications, and mental health monitoring
Comparative Evaluation of Preprocessing Methods for MobileNetV1 and V2 in Waste Classification
Full text linkWaste management remains a critical challenge for many countries, including Indonesia, which ranks as the world's second-largest contributor of waste. As tens of millions of tons are produced each year and the management system remains ineffective, environmental conditions and public health continue to deteriorate. To address this issue, it is imperative to develop more accurate and efficient solutions to enhance waste classification and management. This study investigates the influence of various image preprocessing techniques on the performance of MobileNetV1 and MobileNetV2 models in the classification of waste images. Preprocessing is crucial for enhancing data quality, particularly when dealing with real-world images that are affected by inconsistent lighting, texture, and clarity. Five preprocessing scenarios were evaluated: Baseline, CLAHE with Bilateral Filtering, CLAHE with Sharpening, Grayscale with CLAHE, and Gaussian Blur with Bilateral Filtering. Among these, the combination of CLAHE and Bilateral Filtering applied to MobileNetV1 achieved the best results, with 85% training accuracy, 96% validation accuracy, a training loss of 0.3178, and the lowest validation loss of 0.1630. Overall, MobileNetV1 benefited more significantly from preprocessing variations than MobileNetV2, particularly in terms of accuracy improvement and reduction in prediction error. These findings underscore the importance of effective preprocessing in enhancing model performance for waste image classificatio
Classification of Red Foxes: Logistic Regression and SVM with VGG-16, VGG-19, and Inception V3
Full text linkDeep learning models demonstrate a high degree of accuracy in image classification. The task of distinguishing between various sources of red fox images—such as authentic photographs, game-captured images, hand-drawn illustrations, and AI-generated images—raises important considerations regarding realism, texture, and style. This study conducts an evaluation of three deep learning architectures: Inception V3, VGG-16, and VGG-19, utilizing images of red foxes. The research employs Silhouette Graphs, Multidimensional Scaling (MDS), and t-Distributed Stochastic Neighbor Embedding (t-SNE) to assess clustering and classification efficiency. Support Vector Machines (SVM) and Logistic Regression are utilized to compute the Area Under the Curve (AUC), Classification Accuracy (CA), and Mean Squared Error (MSE). The MDS plots and t-SNE data clearly demonstrate the capability of the three deep learning models to distinguish between the image categories. For game-captured images, VGG-16 and VGG-19 demonstrate quite outstanding performance with silhouette scores of 0.398 and 0.315, respectively. This study explores the enhancement of classification accuracy in logistic regression and support vector machines (SVM) through the refinement of decision boundaries for overlapping categories. Utilizing Inception V3, an artificial intelligence-generated image silhouette score of 0.244 was achieved, demonstrating proficiency in image classification. The research highlights the challenges posed by diverse datasets and the efficacy of deep learning models in the classification of red fox images. The findings suggest that integrating deep learning with machine learning classifiers, such as logistic regression and SVM, may improve classification accuracy