Garuda - Garba Rujukan Digital

Optimization of Numerical Algorithms for Solving Large Linear Equation Systems in Industrial Mathematical Computing M Bastian; Putry Wahyu Setyaningsih; Syeda Azwa Asif
International Journal of Applied Mathematics and Computing Vol. 1 No. 4 (2024): October: International Journal of Applied Mathematics and Computing
Publisher : Asosiasi Riset Ilmu Matematika dan Sains Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.62951/ijamc.v1i4.275

The rapid advancement of modern computing has driven extensive research on numerical algorithms for solving large-scale systems of linear equations. Classical methods such as LU decomposition, Jacobi, and Gauss–Seidel have been revisited and optimized to leverage parallel architectures, GPUs, and even quantum platforms. Recent studies demonstrate that optimized algorithms can reduce computation time by more than 50% while maintaining high accuracy in solving high-dimensional problems. LU decomposition, particularly in its parallel and GPU-based implementations, has shown superior performance in batch processing and industrial-scale simulations. Meanwhile, iterative methods such as Jacobi and Gauss–Seidel remain relevant due to their flexibility in numerical modeling, with further developments for block matrix systems, finite element applications, and FPGA architectures. The integration of these enhanced algorithms is not only beneficial for the advancement of scientific software development but also supports practical applications in engineering simulations, large-scale data optimization, and machine learning. Therefore, an integrative review of modern numerical algorithm developments is crucial in bridging the gap between industrial demands and research progress in scientific computing.

Gamified Android Learning to Foster Higher-Order Thinking in Students with ADHD Susiaty, Utin Desy; Chandra Lesmana
International Journal of Applied Mathematics and Computing Vol. 3 No. 1 (2026): January: International Journal of Applied Mathematics and Computing
Publisher : Asosiasi Riset Ilmu Matematika dan Sains Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.62951/ijamc.v3i1.286

This study addresses the problem of how the implementation of HOTS-based Android gamification influences the higher-order thinking skills of students with ADHD, a group that often faces challenges in traditional learning environments. A quantitative experimental research design was applied, involving 26 students with ADHD from four special needs schools (SLBs) in West Kalimantan. The intervention included HOTS-oriented Android gamified learning, and students' performance was measured using pre-tests and post-tests based on HOTS-level questions. The average pre-test score was 23.72, while the post-test score increased to 53.21. A paired sample t-test showed a significant improvement (t = 8.688 > t_table = 1.708, at a 5% significance level). However, only 57.69% of students met the minimum mastery criteria (KKM = 67), indicating that 15 out of 26 students achieved the expected learning standard. The implementation of HOTS-based Android gamification significantly improved the higher-order thinking skills of students with ADHD. Nonetheless, the overall results, based on average scores and classical completeness, indicate that many students still did not reach the expected level of mastery. Further enhancements in instructional design may be necessary to optimize outcomes for this group of learners.

Comparative Evaluation of YOLOv5, YOLOv7, and YOLOv8 for Outdoor Traffic Object Detection Achmad, Refi Riduan; Abil, Muhammad; Fadhilah, Muhammad Raihan; Sandi
International Journal of Applied Mathematics and Computing Vol. 3 No. 2 (2026): April: International Journal of Applied Mathematics and Computing
Publisher : Asosiasi Riset Ilmu Matematika dan Sains Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.62951/ijamc.v3i2.291

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.

QoS-Based Assessment and Classification of Network Conditions Using OSPF and BGP Routing Protocols Achmad, Refi Riduan; Reza, Muhammad Ali
International Journal of Applied Mathematics and Computing Vol. 3 No. 2 (2026): April: International Journal of Applied Mathematics and Computing
Publisher : Asosiasi Riset Ilmu Matematika dan Sains Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.62951/ijamc.v3i2.293

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.

Identification of Housing Eligibility Status Using Family Data in Samarinda City Antonieta Aryuka Paskalia Nggotu; Hamdani, Hamdani; Anindita Septiarini
International Journal of Applied Mathematics and Computing Vol. 3 No. 2 (2026): April: International Journal of Applied Mathematics and Computing
Publisher : Asosiasi Riset Ilmu Matematika dan Sains Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.62951/ijamc.v3i2.294

The issue of uninhabitable houses still requires an accurate identification mechanism because the manual data collection process has the potential to be time-consuming, costly, and subject to subjectivity in determining aid priorities. This study aims to develop a classification model to identify habitable and uninhabitable houses based on family socioeconomic data using the Random Forest algorithm. The research method includes data preprocessing, data division using stratified split in three scenarios, baseline model development, and optimization through hyperparameter tuning using GridSearchCV with 3-fold cross-validation and balanced class_weight parameters. The data used includes variables such as education type, employment status, occupation type, number of family members, and family insurance type. The test results show that the 70:30 data division scenario after tuning provides the best performance with a recall value of 0.5797 for uninhabitable houses and an F1-score of 0.4746. Feature importance analysis shows that education type and employment status are the most influential variables in the classification. The results of this study show that the model built is capable of increasing sensitivity in detecting uninhabitable houses to support more objective field survey prioritization.

Using the Aquila Optimizer to Estimate the Two Parameters of the Fréchet Distribution Basheer Jameel
International Journal of Applied Mathematics and Computing Vol. 3 No. 2 (2026): April: International Journal of Applied Mathematics and Computing
Publisher : Asosiasi Riset Ilmu Matematika dan Sains Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.62951/ijamc.v3i2.299

The Fréchet distribution is one of the commonly used Extreme Value Distributions (EVDs) in statistical modeling and heavy-tailed data analysis, where it plays an important role in describing product lifetimes as well as climatic and financial phenomena. The estimation of its two parameters, namely the shape parameter and the scale parameter, is traditionally based on the Maximum Likelihood Estimation (MLE) method. However, maximizing the likelihood function for this distribution involves numerical difficulties, which necessitates the use of numerical optimization methods. In this study, we propose the use of the Aquila Optimizer (AO), a recent metaheuristic algorithm inspired by the hunting behavior of eagles, as an efficient numerical tool for maximizing the likelihood function of the Fréchet distribution. The objective function was formulated as the negative log-likelihood function (-LogL), and the Aquila Optimizer was employed to obtain the optimal estimates of the distribution parameters. Several simulation experiments with different sample sizes were conducted to compare the performance of the proposed method with a conventional approach represented by the Nelder–Mead method, using the Mean Squared Error (MSE) criterion. The simulation results demonstrated that the Aquila Optimizer outperformed the Nelder–Mead algorithm in many cases, although the superiority was slight. The results also showed that both algorithms were consistent, as their MSE values decreased with increasing sample size. In addition, a practical application was carried out using real data, and the results of the survival function estimation indicated a good fit.

Combination Of SAW And TOPSIS Methods for Support Decisions on Rice Land Suitability Amalia, Syaffira Rizky; Hamdani, Hamdani; Septiarini, Anindita
International Journal of Applied Mathematics and Computing Vol. 3 No. 2 (2026): April: International Journal of Applied Mathematics and Computing
Publisher : Asosiasi Riset Ilmu Matematika dan Sains Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.62951/ijamc.v3i2.310

Rice plants (Oryza Sativa L.) are the main staple food commodity in Indonesia, as most of the Indonesian population relies on rice as their primary food. One of the causes of low rice production in Indonesia is that farmers generally cultivate rice improperly, such as in land preparation or land selection. Land suitability in rice cultivation greatly affects crop productivity. A process that can support decisions regarding rice land suitability is the development of a Decision Support System (DSS) website using a combination of the Simple Additive Weighting (SAW) method and the Technique for Order Performance of Similarity to Ideal Solution (TOPSIS). This combination is performed by taking the average (µ) of the final results from the SAW and TOPSIS methods. The final scores of each method are calculated separately, and then the average (µ) of these two results is taken to obtain the final ranking of the alternatives. The data used to determine the suitability of rice land is based on five criteria: soil type, soil pH, rainfall, temperature, irrigation and water supply. The alternative data used in the study includes six alternatives: Sungai Kunjang, Sambutan, Samarinda Utara, Palaran, Loa Janan Ilir, and Samarinda Seberang. The aim of this research is to provide information on alternative solutions to farmers or farmer groups in determining rice land suitability. The results of the combination of the SAW and TOPSIS methods show that the alternative with the highest final score is Samarinda Utara (A3), with a final score of 0.7337. Meanwhile, the alternative with the lowest final score is Sambutan (A2), with a final score of 0.4402.

Detecting Levels of Learning Concentration Through Student Behavior in the Classroom Using Convolutional Neural Networks (CNN)
International Journal of Applied Mathematics and Computing Vol. 2 No. 1 (2025): January: International Journal of Applied Mathematics and Computing
Publisher : Asosiasi Riset Ilmu Matematika dan Sains Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.62951/ijamc.v2i1.74

This study discusses a student concentration detection system using Convolutional Neural Network (CNN) with the MobileNetV2 architecture. The dataset was adapted from Classroom Student Behaviors and mapped into four concentration categories: highly focused, focused, less focused, and unfocused. The system was tested with a 720p webcam and produced real-time detection data. The evaluation results show an overall accuracy of 75.85%, with the highest precision achieved in the focused class (0.9859) and the highest recall in the highly focused (0.9739) and unfocused (0.9811) classes. The confusion matrix indicates that the focused class was detected most consistently, while highly focused and unfocused classes were often misclassified as focused, resulting in lower precision. In real-time testing, the system operated at an average of 7 FPS and worked optimally when students faced the camera directly with sufficient lighting, but its performance decreased significantly at face angles greater than 45°. User evaluation shows that 75% of students rated the detection results as accurate/very accurate with an average satisfaction score of 3.6 out of 5, and 75% felt assisted in recognizing their concentration level. From the teachers’ perspective, most stated that the results were consistent with classroom observations, and all expressed willingness to reuse the system.

Prediction of Credit Sales Value with the Naive Bayes Algorithm on Sujase Cell Jakarta
International Journal of Applied Mathematics and Computing Vol. 1 No. 3 (2024): July : International Journal of Applied Mathematics and Computing
Publisher : Asosiasi Riset Ilmu Matematika dan Sains Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.62951/ijamc.v1i3.110

Background: The rapid growth of mobile phone usage has significantly increased the demand for prepaid credit services (mobile airtime), creating large volumes of transaction data that require effective analysis for business decision-making. Sujase Cell, a mobile credit retailer in Jakarta, faces challenges in predicting future sales performance and customer purchasing interest due to the accumulation of transaction records over time and the limitations of manual analysis. Objective: This study aims to identify customer purchasing interest and predict mobile credit sales values by implementing the Naive Bayes algorithm as a data mining approach to support sales forecasting and business development strategies. Methods: The research employed a quantitative predictive approach using a private dataset obtained from Sujase Cell. Data collection was conducted through observation and literature review. The dataset consisted of historical mobile credit sales transactions and sales balance records collected during the study period. The data underwent preprocessing stages, including normalization using the Min-Max Scaler technique, followed by data partitioning into training and testing datasets. The Naive Bayes classification method was then applied to analyze sales patterns and generate predictions. Model performance was evaluated using Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and confusion matrix-based assessment metrics. Several experimental scenarios involving different training-testing ratios and parameter configurations were conducted to determine the most effective predictive model. Results: The findings indicate that the Naive Bayes method successfully identified sales trends and customer purchasing behavior patterns. The best-performing model was obtained using a 90% training dataset and 10% testing dataset, resulting in the lowest prediction error. Experimental results demonstrated that the generated prediction model was capable of following actual sales patterns and producing reliable forecasting outcomes. The implementation of Naive Bayes provides valuable support for sales planning, inventory management, and marketing decision-making at Sujase Cell, enabling the business to improve operational efficiency and anticipate future market demand more effectively.

Data Mining Classification in The New Student Admission Process Using The K-Nearest Neighbors Method : Case Study: Yapmi Boarding School
International Journal of Applied Mathematics and Computing Vol. 1 No. 4 (2024): October: International Journal of Applied Mathematics and Computing
Publisher : Asosiasi Riset Ilmu Matematika dan Sains Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.62951/ijamc.v1i4.111

The new student admission process is an important activity in educational institutions to ensure that prospective students meet the established admission criteria. However, the selection process is often conducted manually, making it less efficient and prone to subjective assessments. This study aims to implement a data mining classification approach using the K-Nearest Neighbors (K-NN) method to support decision-making in the new student admission process at Yapmi Boarding School. The research utilizes historical admission data consisting of academic scores, interview results, and other admission criteria as classification attributes. The K-NN algorithm was applied to classify prospective students into accepted and rejected categories based on the similarity of their characteristics to previously evaluated applicants. The research methodology includes data collection, preprocessing, classification modeling, and performance evaluation using accuracy metrics. The results demonstrate that the K-NN method is capable of classifying prospective students effectively and can assist admission committees in making more objective and accurate decisions. The implementation of this model contributes to improving the efficiency, consistency, and reliability of the student admission process at Yapmi Boarding School. Therefore, the K-NN algorithm can be considered a viable alternative for supporting educational admission decision systems.

Home Page

OAI Link

Editorial Team

Contact

Reviewer

Google Scholar

Contact Name
Rusliadi

Contact Email
garuda@apji.org

Phone
+6285642100292

Journal Mail Official
fatqurizki@apji.org

Editorial Address
Jln. Perum Cluster G11 Nomor 17 Jl. Plamongan Indah, Pedurungan, Semarang, Provinsi Jawa Tengah, 50195

Location
Kota semarang,

Jawa tengah

INDONESIA

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Home Page

OAI Link

Editorial Team

Contact

Reviewer

Google Scholar

Contact Name Rusliadi

Contact Email garuda@apji.org

Phone +6285642100292

Journal Mail Official fatqurizki@apji.org

Editorial Address Jln. Perum Cluster G11 Nomor 17 Jl. Plamongan Indah, Pedurungan, Semarang, Provinsi Jawa Tengah, 50195

Location Kota semarang, Jawa tengah INDONESIA

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Contact Name
Rusliadi

Contact Email
garuda@apji.org

Phone
+6285642100292

Journal Mail Official
fatqurizki@apji.org

Editorial Address
Jln. Perum Cluster G11 Nomor 17 Jl. Plamongan Indah, Pedurungan, Semarang, Provinsi Jawa Tengah, 50195

Location
Kota semarang,

Jawa tengah

INDONESIA