Publication Search

Object detection plays a crucial role in intelligent transportation systems, particularly for outdoor traffic monitoring applications that require accurate and real-time performance under limited computational resources. Recent developments in YOLO-based architectures have introduced multiple model variants; however, their practical performance under constrained training conditions remains insufficiently explored. This study presents a comparative evaluation of YOLOv5, YOLOv7, and YOLOv8 for outdoor traffic object detection using a real-world dataset and identical experimental settings. The main objective of this research is to analyze the robustness and detection quality of different YOLO variants when trained with a limited number of epochs, reflecting practical deployment scenarios. All models were trained and evaluated using the same dataset, preprocessing pipeline, and hardware configuration to ensure a fair comparison. Performance evaluation was conducted using multiple metrics, including precision, recall, mAP@50, Precision–Recall curves, area under the curve (AUC), and peak F1-score. Experimental results indicate that YOLOv5 outperformed YOLOv7 and YOLOv8 in terms of overall detection stability and robustness. The merged Precision–Recall analysis shows that YOLOv5 achieved a higher effective AUC and superior mAP@50, reflecting better global detection performance. In addition, YOLOv5 exhibited a higher peak F1-score, indicating a more balanced trade-off between precision and recall. In contrast, YOLOv7 and YOLOv8 showed performance degradation under limited training conditions despite their more advanced architectures. These findings suggest that YOLOv5 remains a reliable and efficient solution for outdoor traffic object detection, particularly in resource-constrained environments. The study highlights the importance of comprehensive evaluation metrics and practical experimental settings when selecting object detection models for real-world applications.

Object detection plays a crucial role in intelligent transportation systems, particularly for outdoor traffic monitoring applications that require accurate and real-time performance under limited computational resources. Recent developments in YOLO-based architectures have introduced multiple model variants; however, their practical performance under constrained training conditions remains insufficiently explored. This study presents a comparative evaluation of YOLOv5, YOLOv7, and YOLOv8 for outdoor traffic object detection using a real-world dataset and identical experimental settings. The main objective of this research is to analyze the robustness and detection quality of different YOLO variants when trained with a limited number of epochs, reflecting practical deployment scenarios. All models were trained and evaluated using the same dataset, preprocessing pipeline, and hardware configuration to ensure a fair comparison. Performance evaluation was conducted using multiple metrics, including precision, recall, mAP@50, Precision–Recall curves, area under the curve (AUC), and peak F1-score. Experimental results indicate that YOLOv5 outperformed YOLOv7 and YOLOv8 in terms of overall detection stability and robustness. The merged Precision–Recall analysis shows that YOLOv5 achieved a higher effective AUC and superior mAP@50, reflecting better global detection performance. In addition, YOLOv5 exhibited a higher peak F1-score, indicating a more balanced trade-off between precision and recall. In contrast, YOLOv7 and YOLOv8 showed performance degradation under limited training conditions despite their more advanced architectures. These findings suggest that YOLOv5 remains a reliable and efficient solution for outdoor traffic object detection, particularly in resource-constrained environments. The study highlights the importance of comprehensive evaluation metrics and practical experimental settings when selecting object detection models for real-world applications.

This study aims to develop a student learning progress monitoring system designed to assist teachers and parents in tracking students’ academic performance at SMP Negeri 2 Loura. The main issue identified in the school is that academic information is still distributed using manual procedures, which results in delays and limited transparency regarding students’ learning progress. To address this problem, the system was developed using the Model View Controller (MVC) architecture and the Waterfall approach to system development, which consists of several sequential phases such as analyzing requirements, designing the system, implementing the solution, conducting tests, and performing ongoing maintenance. The findings of this research show that the system that has been created is capable of presenting academic information in a complete and structured manner, including assignment scores, daily tests, and semester examinations. The system provides faster and easier access for teachers to input grades and for parents to monitor their children’s academic development in real time. Functional testing shows that all features operate correctly according to user needs, with no errors found during system operation.

This study aims to analyze the effect of welfare inequality on poverty levels on the island of Sumatra. Welfare inequality is measured using the Gini Index, while poverty levels are measured by the percentage of the poor population at the provincial level. This study uses a quantitative method with a panel data approach covering ten provinces on the island of Sumatra during the period 2020–2024. The analytical techniques used include panel data regression with fixed and random effects models, as well as classical assumption testing to ensure model validity. The results show that welfare inequality has a positive and significant effect on poverty levels, meaning that increasing inequality in income distribution tends to increase the number of poor people. This finding indicates that uneven economic growth can worsen the welfare of the community, especially low-income groups. Therefore, more inclusive and sustainable development policies are needed, particularly in increasing equitable access to education, health services, and productive employment opportunities to reduce inequality and poverty levels on the island of Sumatra.

This research aims to develop a digital-based governance model for village finance to support the realization of a Smart and Green Village. The study analyzes how digital transformation in village finance management can increase transparency, accountability, and efficiency, while promoting sustainability-oriented budgeting and environmental practices. A Systematic Literature Review (SLR) was used to identify, evaluate, and synthesize scientific publications from 2022 to 2025, accessed through reputable databases such as ScienceDirect, Springer, Wiley, Taylor & Francis, SAGE, ACM, and IEEE. The review focused on topics including digital governance in villages, digital public finance, smart village development, green budgeting, environmental sustainability, and rural digital transformation. Findings indicate that digital-based village finance governance enhances administrative efficiency, strengthens budget transparency through real-time monitoring, minimizes financial deviation risks, and boosts public participation in fiscal accountability. Integrating digital systems with green budgeting features enables the prioritization of sustainable programs, such as renewable energy, waste management, climate change mitigation, and green infrastructure development. The study suggests a comprehensive digital-based governance model that includes e-budgeting, e-accounting, digital payment systems, public transparency dashboards, and environmental performance indicators to support the implementation of a Smart and Green Village. This research offers strategic insights for village governments, policymakers, and practitioners on the importance of adopting digital governance tools to improve financial management and strengthen sustainable development at the local level.

This research, entitled “Implementation of a Web-Based Student Tuition Payment Information System at SMA Swasta Manda Elu,” aims to improve the efficiency of recording and managing students’ tuition fee payments. Previously, the tuition payment process was handled manually, which often led to recording errors, delays in reporting, and difficulties in retrieving student payment data. To address these issues, a web-based information system was developed to assist the school administration in recording, reporting, and monitoring tuition payments more effectively and efficiently. The development method used in this study is the Model View Controller (MVC) approach, which separates the system’s view, logic, and database layers to simplify system management and maintenance. The final result of this research is a web-based tuition payment application that runs properly, provides convenience for school administrative staff in recording transactions, and enhances the accuracy and speed of the school’s financial administration process at SMA Swasta Manda Elu.

This study investigates the impact of household consumption and investment on the open unemployment rate in West Kalimantan over both short and long terms. Using time series data from 1995–2024 and employing the Vector Error Correction Model (VECM) in EViews 12, results show all variables become stationary after first differencing. Johansen cointegration confirms a long-run equilibrium relationship. VECM findings reveal that investment significantly increases unemployment in the long run, while household consumption has no significant effect. In the short term, a significant error correction mechanism exists, indicating adjustment toward long-run equilibrium. Diagnostic tests confirm model validity through absence of autocorrelation and normally distributed residuals. These results highlight the need for more targeted policies to reduce unemployment.

This study aims to analyze the causal relationship between interest rates, exchange rates, and inflation in Indonesia during the period 1994–2023 using the Vector Error Correction Model (VECM) approach. The data used are monthly time series secondary data obtained from the World Bank. The Johansen cointegration test results indicate a long-term relationship among the three variables. However, the Granger causality test finds no significant short-term causal relationship. The VECM estimation reveals that inflation is the most responsive variable in correcting long-term disequilibrium, while the exchange rate plays a dominant role in influencing both inflation and interest rates. The Impulse Response Function and Variance Decomposition results indicate that these variables interact dynamically, especially in the medium to long term. These findings highlight the importance of exchange rate stabilization and enhancing the effectiveness of monetary policy to maintain macroeconomic stability in Indonesia.

This study aims to examine the effect of Return on Assets (ROA) and Return on Equity (ROE) on stock prices of agro-industrial companies listed on the Indonesia Stock Exchange (IDX) during the post-Covid-19 economic recovery period, specifically from 2021 to 2023. The research method used is a quantitative approach with a causal associative type. The data used are secondary data from annual financial statements and year-end stock prices. Data analysis was conducted using multiple linear regression with classical assumption tests to ensure model validity. The results indicate that ROA and ROE have both simultaneous and partial effects on stock prices. ROA demonstrates a stronger influence compared to ROE, suggesting that asset management efficiency is a key determinant of stock value in agro-industrial firms during the economic recovery period.

This research investigates the implementation of Random Forest algorithms combined with Synthetic Minority Over-sampling Technique (SMOTE) to predict elementary school dropout rates in Indonesia, supporting the Indonesia Emas 2045 vision. A significant gap was identified in previous studies, which, despite utilizing artificial intelligence for dropout interventions, had not integrated temporal dimensions into data analysis. A temporal data-based classification model was developed using Indonesian Ministry of Education data from 2021-2023, incorporating lag features, delta calculations, and rolling statistics. Two models were implemented: one with SMOTE achieving 99% accuracy with perfect recall for high-risk regions, while the non-SMOTE model reached 100% accuracy. Temporal features were identified as crucial predictors, reflecting external fluctuations and annual changes impacting dropout decisions. This approach enables educational institutions to allocate resources more efficiently by prioritizing operational assistance for high-risk schools. The model's capacity to identify high-risk regions with 100% recall represents a strategic investment in strengthening Indonesia's human resource sustainability. To address the limitations of provincial aggregate data, expansion to include individual-level variables and model validation at district or school scales is recommended for future research.

The stability of the banking sector is crucial in maintaining a country’s financial system and economic sustainability. This study analyzes the impact of Net Interest Margin (NIM), Non-Performing Loans (NPL), and Capital Adequacy Ratio (CAR) on banking stability in Indonesia. The inconsistency of previous research findings indicates a research gap that requires further exploration. This study employs a quantitative approach using secondary data from financial reports of conventional banks listed on the Indonesia Stock Exchange (IDX) from 2019 to 2023. The sample was selected using purposive sampling, resulting in 39 banks with 195 observations. Data analysis was conducted using multiple linear regression with classical assumption tests, including normality, heteroscedasticity, multicollinearity, and autocorrelation, to ensure model validity. The results show that NIM positively but not significantly affects banking stability, while NPL has a negative and significant effect. CAR also significantly influences banking stability. Enhancing banking intermediation effectiveness through NIM and CAR can strengthen financial stability, whereas increasing credit risk, reflected in NPL, can weaken stability. This study provides insights for regulators and banking management in designing more effective policies to maintain banking sector stability in Indonesia.

Penelitian ini bertujuan untuk mengeksplorasi faktor makroekonomi yang mempengaruhi ekspor di Indonesia. Data yang digunakan dalam penelitian ini adalah data sekunder yaitu data ekspor, inflasi, suku bunga, dan nilai tukar pada periode 1998-2022 yang dipublikasikan oleh Badan Pusat Statistik dan Bank Indonesia. Penelitian ini menggunakan pendekatan kuantitatif. Metode analisis data yang digunakan dalam penelitian ini adalah model Vector Error Correction Model (VECM) dengan data time series, data diolah dengan menggunakan program eviews 10. Hasil pengujian VECM dalam jangka panjang inflasi berpengaruh positif dan signifikan terhadap ekspor, suku bunga berpengaruh negatif dan signifikan terhadap ekspor. Untuk jangka pendek inflasi berpengaruh negatif dan signifikan terhadap ekspor, suku bunga berpengaruh positif dan signifikan terhadap ekspor, dan nilai tukar pada jangka panjang dan pendek berpengaruh negatif dan tidak signifikan terhadap ekspor.

Damage that occurs to manufactured goods can occur due to several things, such as in the manufacturing process, in the packaging process or during the delivery process. To find out where when the goods were damaged and the factors causing the damage to the goods as well as the efforts made to reduce the damage to the goods, an investigation was carried out at the time of unloading the goods, to find out where the damage to the goods occurred so as not to harm the recipient of the goods if they received the goods in damage condition.Based on multiple linear regression analysis, the results obtained are Y = 3.133 + 0.044 X1, + 0.402 X2 + 0.299 X3 + . Partial test results of machine (X1), material (X2), man power (X3) have a positive and significant effect on damage goods. It is proven by the results of the comparison of the value of t count with t table and with a significant level comparison of 5% (0.05) where the machine variable t count (0.413) < t table (1.98609), material variable t count (3.142) > t table (1.98609), variable man power t count (2.990) > t table (1.98609). The R2 test of machines, materials and man power gave a significant effect on damage goods by 43.6%, while the remaining 56.4% was explained by other causes outside the model variables that were not examined.

In order to better understand the complex links between infrastructure development, education, tourism, agriculture, and regional economic growth in Indonesia, this study uses structural equation modeling with partial least squares, or SEM-PLS. The research makes use of a 287-observation dataset and applies rigorous measurement model evaluation, bootstrapping analysis, structural model estimation, and model validation against multiple fit indices. The results show a strong positive correlation between each sector—agriculture, tourism, education, and infrastructure development—and regional economic expansion. The necessity for comprehensive, integrated development strategies is highlighted by the interconnectedness of these interactions. In order to promote balanced and resilient regional economic growth, the findings provide policymakers with useful insights and highlight the significance of ongoing investments in infrastructure, education, and sustainable tourism and agricultural practices.    

The application of the VARK (Visual, Auditory, Reading/Writing, Kinesthetic) model in classroom management is a strategy that can offer a more inclusive and effective approach to increasing student achievement. This research aims to explore how the use of the VARK model can improve students' learning experiences and academic performance. The research method used is a literature study, involving analysis of various relevant sources regarding the VARK concept and its application in education. The research results show that by understanding various learning preferences, educators can design learning experiences that better suit students' individual needs. Through the application of various learning strategies, such as the use of pictures and diagrams for visual students, group discussions for auditory students, reading and writing assignments for students who prefer to learn through text, as well as practical activities or experiments for kinesthetic students, students can be more involved in learning and gain a deeper understanding. This research also emphasizes the importance of flexibility in teaching, adapting to changes in student learning preferences, and evaluating the effectiveness of the learning strategies used. Although the VARK model has the advantage of providing guidance for inclusive teaching, it is important to remember that students' learning preferences may change and that various factors also influence their academic performance.

This research aims to determine the effect of the Problem Based Learning (PBL) learning model on student learning outcomes in the Biogeochemical Cycle sub-material in Class X SMA Negeri 1 Darangdan. This research uses a quantitative approach with a Quasi Experimental research method with a Non Equivalent Control Group Design. The population in this study were all class X students of SMA Negeri 1 Darangdan, Purwakarta Regency. The sampling technique was Purposive Sampling with a sample of class X.9 as the experimental class which took part in learning by applying the Problem Based Learning (PBL) learning model and class The variables of this research include the independent variable, namely the Learning Model (PBL Learning Model vs. Conventional Learning Model) and the dependent variable, namely student learning outcomes in the Biogeochemical Cycle sub-material in Class X SMA Negeri 1 Darangdan. The data collection technique uses the Test Method. The instrument used is a multiple choice test. Learning outcome data were analyzed using the Independent Sample T-test. Learning Results of experimental class students achieved good characteristics and the control class only achieved adequate characteristics. The mean post-test for the experimental class was 83.1 and the mean post-test for the control class was 74.04. The results of the Independent Sample t-test show the value of tcount = 3.363 > ttable = 1.989. Because tcount > ttable, Ho is rejected and H1 is accepted, which means that there is a significant difference in student learning outcomes between classes that take part in learning using the Problem Based Learning (PBL) learning model and classes that take part in conventional learning. And also the Sig value. (2-tailed) < 0.05, namely 0.001. So it can be concluded that the Problem Based Learning (PBL) learning model influences student learning outcomes.      

Fraudsters increasingly exploit unauthorized credit card information for financial gain, targeting un-suspecting users, especially as financial institutions expand their services to semi-urban and rural areas. This, in turn, has continued to ripple across society, causing huge financial losses and lowering user trust implications for all cardholders. Thus, banks cum financial institutions are today poised to implement fraud detection schemes. Five algorithms were trained with and without the application of the Synthetic Minority Over-sampling Technique (SMOTE) to assess their performance. These algorithms included Random Forest (RF), K-Nearest Neighbors (KNN), Naïve Bayes (NB), Support Vector Machines (SVM), and Logistic Regression (LR). The methodology was implemented and tested through an API using Flask and Streamlit in Python. Before applying SMOTE, the RF classifier outperformed the others with an accuracy of 0.9802, while the accuracies for LR, KNN, NB, and SVM were 0.9219, 0.9435, 0.9508, and 0.9008, respectively. Conversely, after the application of SMOTE, RF achieved a prediction accuracy of 0.9919, whereas LR, KNN, NB, and SVM attained accuracies of 0.9805, 0.9210, 0.9125, and 0.8145, respectively. These results highlight the effectiveness of combining RF with SMOTE to enhance prediction accuracy in credit card fraud detection.

Clean architecture is a method of application development that divides code into multiple layers based on the purpose of the code, ensuring minimal dependency. Popular architectures in Android application development include MVP (Model View Presenter), MVI (Model View Intent), and MVVM (Model View ViewModel). This research focuses on creating three applications with similar features and interfaces using different architectures. The study compares modifiability, testability, and performance aspects to determine the differences between each architecture. The results show that MVVM architecture is the best in modifiability, with the lowest number of index modifications. Testability requires no more than four scenarios for all architectures. However, MVI architecture outperforms in test coverage, and MVP architecture outperforms in performance. Overall, clean architecture is a valuable approach for improving the performance and usability of Android applications.

This study is entitled to determine the effect of Work Discipline, Work environment and Leadership Models on Work Loyalty. In this research using quantitative methods. The sample in this study were Proseecing employees at PT Syngenta Seed Indonesia PIER, Pasuruan Regency with a sample size of 106 respondents who were taken using a purposive sampling saturated. The analytical method used is multiple linear regression analysis. To obtain the test results in this study, validity and reliability tests were carried out, classical assumption tests were normality and heroscedasticity tests, and then hypothesis testing was carried out. The results of this study show that the results of the Determination Test (R2) for the variable Work Motivation, Work Discipline, Work Environment and Leadership Model have an effect on Work Loyalty of 0.446 or the equivalent percentage value of 46.3%, while the remaining 55.4% is influenced by other variables . For the significant level, namely the the Work Environment variable has a nt significant effect  Work Loyalty, while the Work Discipline variable has significant effect on Work Loyalty, and the Leadership Model variable has significant effect on Work Loyalty,

The purpose of the study is to produce a learning model with valid, practical, and effective criteria. This type of research is development research using the Borg And Gall model. The population is all grade V students of State Elementary School 18 West Freshwater, Padang Utara District, Padang City. The sample was 30 class V students. The sampling technique is total sampling.  The model validation design is validated by 2 experts: motor and language and model practicality by 2 sports teachers, The data collection instrument consists of tests of validity, practicality and effectiveness. As well as data analysis techniques using descriptive qualitative and quantitative techniques. The validation results of the learning model were obtained on average by 81.2.% with valid categories. In practicality, the results were obtained that this learning model was practical or good for use with learning implementation results of 91.7% The results of the effectiveness test using the Manipulative Basic Motion learning model based on play activities showed that there was a difference between obtaining manipulative motion ability results before using the model and after using the learning model. It was concluded that the model is effective and can improve the results of manipulative motion abilities (control objects) in the learning of sports and health physical education in grade V of primary school.