ejournal.nusamandiri.ac.id (STMIK Nusa Mandiri)
Not a member yet
1504 research outputs found
Sort by
FACE DETECTION FOR ENTERPRISE RESOURCE PLATFORM ATTENDANCE SYSTEM: A COMPARATIVE ANALYSIS
Get PDFThe demand for face detection capabilities in attendance systems has led to various implementations using different algorithms and Enterprise Resource Planning (ERP) platforms. This research aimed to conduct a comparative analysis of three face detection algorithms—Dlib, Haar-Cascade, and MTCNN (Multi-task Cascaded Convolutional Networks)—and implement the most effective solution in an Odoo-based attendance system supporting multiple face detection. The study employed evaluation methodology analyzing real-time video streams, utilizing distinct datasets: a control dataset under standard conditions and a challenge dataset featuring variations in lighting, occlusions, and multiple simultaneous faces. Performance evaluation was based on true positive, false positive, and false negative rates for face detection across both datasets. Results demonstrated significant performance variations: under controller conditions, MTCNN achieved 99.69% detection accuracy compared to Dlib’s 92.92% and Haar-Cascade’s 84.08%, while in challenging environments, MTCNN maintained 60.93% accuracy versus Dlib’s 0.66% and Haar-Cascade’s 2.36%. The significant performance drop in challenging conditions can be attributed to poor lightning conditions, facial occlusions, and the complexity of detecting multiple faces simultaneously. The findings facilitated the development of a custom Odoo attendance module implementing MTCNN, demonstrating potential for improving automated attendance efficiency in organizations while establishing benchmarks for futher development of face recognition-based features within Odoo ERP
IMPLEMENTASI METODE EXTREME MENGGUNAKAN ALGORITMA MACHINE LEARNING UNTUK KLASIFIKASI JENIS FASHION ALAS KAKI
Get PDFThis research explores the application of Extreme Learning Machine (ELM) for classifying types of fashion footwear. The increasing number of e-commerce transactions and the use of visual media in marketing demand an efficient and accurate automated system to identify various types of footwear such as shoes, sandals, and slip-ons. Conventional classification systems often encounter challenges in handling variations in shape, color, and lighting conditions in footwear images. ELM, with its unique approach of assigning random weights in the hidden layer, offers a potential solution to these issues. In this study, a classification system was developed consisting of several stages, including the collection of diverse footwear image data, image preprocessing to improve quality and reduce noise, feature extraction relevant for distinguishing footwear types, and finally, classification using the ELM algorithm. The preprocessing process involved color conversion from RGB to HSV to reduce sensitivity to lighting variations, as well as thresholding to produce binary images. Extracted features included geometric characteristics such as area, perimeter, and aspect ratio. The system’s performance was evaluated using standard metrics such as accuracy, precision, and recall. The results showed an accuracy value of 83.3%. In addition, the model evaluation demonstrated very good results: precision reached 83.3%, recall 83.3%, and F1-Score 91%, indicating that ELM is effective in classifying types of fashion footwear. This study contributes to the development of intelligent, efficient, and accurate classification systems for applications in the fashion industry, while also opening opportunities for further research in optimizing ELM parameters and exploring more representative feature
PENINGKATAN AKURASI KNN DALAM PREDIKSI KELULUSAN MAHASISWA MELALUI OPTIMASI PARAMETER PSO
Get PDFPredicting student graduation is a crucial aspect in supporting academic planning and ensuring timely completion of studies. However, no prior research has specifically applied the integration of K-Nearest Neighbor (KNN) and Particle Swarm Optimization (PSO) for graduation prediction using student data. This study aims to evaluate the effectiveness of combining KNN and PSO in improving classification accuracy. The KNN algorithm is used for classification, while PSO is implemented as a feature selection technique to identify the most relevant attributes. A dataset of 750 student records was processed through data preprocessing and attribute weighting using PSO, followed by model training and evaluation with 10-fold cross-validation. The evaluation results show that the KNN+PSO model improves accuracy from 80.91% to 84.31%, along with increases in precision and recall. These findings indicate that PSO enhances the performance of KNN, particularly in identifying students likely to graduate on tim
PELATIHAN KOMPETENSI PEDAGOGIK GURU DALAM PEMBELAJARAN LITERASI STEAM PADA KURIKULUM MERDEKA DI KEPULAUAN SERIBU
Get PDFEducators in Kepulauan Seribu District, DKI Jakarta, experience challenges in designing and implementing an independent curriculum in Early Childhood Education (ECE), especially in project learning that includes elements of religion, character, identity, literacy, and STEAM. In addition, the lack of curriculum documents in PAUD prompted the need for assistance developing an effective Education Unit Operational Curriculum (KOSP). Through participatory training and mentoring, focusing on improving pedagogical competencies, this service program assists PAUD teachers in preparing learning documents, implementing effective learning, and conducting assessments. The methods used include participatory training, focus discussions, and workshops, encouraging participants to more actively understand, analyze, and apply the STEAM Literacy model in the Merdeka Curriculum. The results of this activity show a significant increase in teacher competence, where the pre-test results show the categories of excellent 8.1%, good 13.5%, sufficient 21.6%, and less 56.8%, while the post-test increased to excellent 48.6%, good 27.1%, sufficient 18.9%, and less 5.4%. The program also provides illustrative examples of concepts, practice videos, and the use of a fun L-STEAM project learning model. Children were more active in learning with loose parts media and natural materials, as evidenced by their enthusiasm during project-based play, despite being supervised by many teachers around them. This program improved PAUD teachers' understanding of curriculum development, learning implementation, and early childhood assessment
Pengaruh Kualitas Produk, Kualitas Pelayanan, dan Influencer Marketing terhadap Kepuasan Pelanggan pada Sego Sambel Merdeka
Get PDFIndonesia’s culinary scenes today has grown and expanded significantly in each region. Increasing culinary demand affects on customer satisfaction. The study’s purpose is to ascertain impact of product quality, service quality, and influencer marketing on customer satisfaction of Sego Sambel Merdeka on Mojokerto which is secured to raise sales, share of the market, and competitive edge. The approach of this study is quantitative by using respondent questionnaires using convenience sampling of 50 respondents. The study’s hypothesis test used the method of multivariate linear regression. The findings of this study indicates results of the validity and realibility tests which have been processed are valid and reliable. The traditional assumption test in this study revealed a normal distribution and no symptoms of multicollinearity and heteroskedasticity occurred. The determination test revealed that 41.2% of the variability of customer satisfaction was impacted by product quality, service quality, and influencer marketing, with a regression equation of Y = -1.689 + 0.336X1 + 0.190X2 + 0.771X3 + e which showed a positive influence. The conclusion obtained on this research is that product quality and influencer marketing have significantly impact on customer satisfaction, but service quality have not significantly impact on customer satisfaction, so in this research, it is implied that Sego Sambel Merdeka business owners are advised to focus more on improving product quality and maximizing influencer marketing strategies, even though the quality of service is not significant in this study periodic evaluation to ensure adequate service standards.Dunia kuliner di Indonesia saat ini telah bertumbuh kembang secara signifikan dan meluas di tiap-tiap daerah. Permintaan kuliner yang meningkat berdampak pada kepuasan pelanggan. Penelitian ini bertujuan untuk mengetahui pengaruh kualitas produk, kualitas pelayanan, dan influencer marketing terhadap kepuasan pelanggan Sego Sambel Merdeka di Mojokerto yang dimana untuk meningkatkan penjualan, pangsa pasar, serta keunggulan kompetitif. Pendekatan penelitian ini bersifat kuantitatif dengan menggunakan pengambilan kuesioner responden dengan menggunakan convenience sampling sebanyak 50 responden. Pengujian hipotesis dalam penelitian ini menggunakan metode regresi linier berganda. Hasil penelitian ini menunjukkan bahwa hasil uji validitas dan realibilitas yang telah diolah valid dan andal. Uji asumsi klasik pada penelitian ini menunjukkan distribusi normal dan tidak terjadi gejala multikolinearitas dan heterokedastisitas. Uji determinasi mengungkapkan bahwa 41,2% variabilitas kepuasan pelanggan dipengaruhi kualitas produk, kualitas pelayanan, dan influencer marketing, dengan persamaan regresi Y = -1,689 + 0,336X1 + 0,190X2 + 0,771X3 + e yang memperlihatkan pengaruh positif. Kesimpulan yang didapat pada penelitian ini bahwa kualitas produk dan influencer marketing berpengaruh signifikan terhadap kepuasan pelanggan, namun kualitas pelayanan tidak berpengaruh signifikan terhadap kepuasan pelanggan
THE ROLE OF L1 REGULARIZATION IN ENHANCING LOGISTIC REGRESSION FOR EGG PRODUCTION PREDICTION
Get PDFPoultry egg productivity is strongly influenced by various environmental factors, such as air and water quality. However, accurately predicting productivity remains a challenge due to the complex interplay of multiple environmental variables and the risk of overfitting in predictive models. This study improves egg productivity prediction using Logistic Regression with L1 regularization, which enhances model generalization by performing automatic feature selection. The research methodology includes data collection of key environmental indicators—Air Quality Index (AQI), Water Quality Index (WQI), and Humidex—followed by data preprocessing, exploratory data analysis (EDA), and model training using L1-regularized Logistic Regression. Model evaluation was performed using classification metrics and learning curve analysis to assess stability and effectiveness. Experimental results indicate that Logistic Regression without regularization achieved an accuracy of 90.7%, with misclassification occurring in the lower production categories. By applying L1 regularization, accuracy increased significantly to 97%, demonstrating its ability to reduce overfitting while improving classification performance. This study also compares its findings with previous research, such as De Col et al. (wheat epidemic prediction, 80–85% accuracy) and Jia Q1 et al. (heart disease prediction, 92.39% accuracy), confirming that our approach outperforms prior Logistic Regression models in similar predictive tasks. These findings suggest that L1 regularization is an effective solution for improving egg productivity prediction, particularly in scenarios with complex environmental influences. Future work will explore advanced ensemble learning techniques and real-time IoT-based monitoring to further enhance prediction accuracy and practical applicability
ZTSCAN: ENHANCING ZERO TRUST RESOURCE DISCOVERY WITH MASSCAN AND NMAP INTEGRATION
Get PDFImplementing Zero Trust Architecture (ZTA) requires a comprehensive understanding of network assets as a fundamental step in implementing security policies. This study proposes ZTscan, an automated tool to increase the efficiency of network asset resource discovery. This proposed tool is then made open source in Github for anyone to evaluate and extend. The research constructs a GNS3-based testing scenario to evaluate the performance of the proposed tool against other scanning tools, including standalone Nmap, Masscan, RustScan, and ZMap. The evaluation focuses on three key metrics: accuracy, scanning speed, and generated data throughput. Experimental results demonstrate that ZTscan achieves 100% accuracy, matching Nmap_Pingsyn while outperforming faster tools such as Masscan, ZMap, and RustScan in precision. ZTscan completes scans 10.64%, faster than Nmap TCP SYN scan while maintaining comparable high accuracy. In terms of throughput, ZTscan reaches a stable peak throughput that is 13.8% lower than Nmap TCP SYN scan without causing disruptive traffic spikes. The findings of this study serve as a reference for resource discovery strategies in ZTA implementation, particularly in scenarios that require fast and accurate network scanning while minimizing potential disruptions or network interference
PEMANFAATAN TEKNOLOGI INFORMASI UNTUK DESAIN KEMASAN DAN LABEL SARUNG GEBENG DESA LIMBANG JAYA
Get PDFGebeng Sarong is a typical woven fabric from Limbang Jaya village, Ogan Ilir Regency, South Sumatra. This craft is increasingly disappearing due to the difficulty of marketing carried out by Limbang Jaya village craftsmen. The packaging and brands used on Gebeng Sarongs are still very simple and unattractive so their marketing is less able to compete with similar products. The lack of creativity of craftsmen in Limbang Jaya is also due to the low level of education of the craftsmen, only elementary or junior high school graduates. In addition, the location of the village is rather remote, causing Gebeng Sarongs to have limited marketing. To help overcome this problem, the Sriwijaya University team held a training activity for craftsmen to develop their products for a wider market with attractive packaging. The stages of this activity include: problem identification, activity preparation, activity implementation, and activity evaluation. The activities carried out were in the form of training activities for craftsmen on the use of information technology and graphic applications., Training activity materials included an introduction to the Canva application, tutorials on making packaging and packaging designs, and direct practice. Based on the results of the pre-test and post-test it can be concluded that this activity has succeeded in improving the skills of craftsmen in designing aesthetic packaging according to market needs, thereby increasing the appeal of Gebeng Sarong in a wider market. This activity is also expected to help maintain the Gebeng Sarong as the typical fabric of the South Sumatra
OPTIMIZING SHUFFLENET WITH GRIDSEARCHCV FOR GEOSPATIAL DISASTER MAPPING IN INDONESIA
Get PDFAccurate classification of natural disasters is crucial for timely response and effective mitigation. However, conventional approaches often suffer from inefficiency and limited reliability, highlighting the need for automated deep learning solutions. This study proposes an optimized Convolutional Neural Network (CNN) based on the lightweight ShuffleNet architecture, enhanced through GridSearchCV for systematic hyperparameter tuning. Using a geospatial dataset of 3,667 images representing earthquake, flood, and wind-related disasters in Indonesia, the optimized ShuffleNet model achieved a peak accuracy of 99.97%, outperforming baseline CNNs such as MobileNet, GoogleNet, ResNet, DenseNet, and standard ShuffleNet. While these results demonstrate the potential of combining lightweight architectures with automated optimization, the exceptionally high performance also indicates possible risks of overfitting and dataset bias due to limited variability. Therefore, future research should validate this approach using larger, multi-source datasets to ensure robustness and real-world applicabilit
ENHANCING COFFEE PRODUCTION FACTOR ASSESSMENT USING LINEAR REGRESSION AND AHP FOR RELIABLE WEIGHT CONSISTENCY
Get PDFThe agricultural sector, particularly coffee production, plays a crucial role in Indonesia’s economy as both a domestic commodity and an export product. However, efforts to optimize coffee production are often constrained by traditional Multi-Criteria Decision-Making (MCDM) methods that rely heavily on subjective judgments, leading to potential inconsistencies—especially in the presence of multicollinearity among variables. This study addresses that challenge by proposing a data-driven and consistent weighting method that integrates Multiple Linear Regression (MLR) with the Analytic Hierarchy Process (AHP). Regression coefficients derived from MLR—based on variables such as the area of immature (-0.2419), mature (0.8357), and damaged (0.5119) coffee plantations—are normalized and incorporated into the AHP pairwise comparison matrix. The resulting Consistency Ratio (CR) values are all below 0.1, indicating high internal consistency and statistical reliability of the derived weights. This integrated approach offers an objective and systematic foundation for MCDM in coffee production analysis, enhances the accuracy of agricultural planning, and supports evidence-based policymaking, while also providing a replicable model for addressing similar challenges in other sector