Jurnal Multimedia dan Teknologi Informasi (Jatilima)

Round-Trip Time Estimation Using Hybrid Neuro-Fuzzy Based On Subtractive Clustering

2026-03-18T17:43:03+00:00

Round-trip time (RTT) is a key latency indicator for quality-of-service (QoS) control and task orchestration in cloud–edge systems. However, RTT is highly time-varying due to congestion dynamics, routing changes, and fluctuating traffic conditions, motivating short-term prediction to enable proactive decision making. This paper investigates a hybrid neuro-fuzzy baseline for RTT prediction implemented using an Adaptive Neuro-Fuzzy Inference System (ANFIS) with subtractive-clustering-based initialization to avoid rule explosion in high-dimensional inputs. A controlled dataset was generated in Mininet using a dumbbell topology with injected delays (1–1000 ms). In total, 100,000 raw RTT records were collected (100 RTT measurements per run across 1000 runs) and aggregated into 1000 supervised samples paired with TCP-state features. Experiments followed a unified and reproducible protocol with a fixed 60/10/30 train/validation/test split, train-only feature standardization, train-only target normalization with inverse transformation for reporting, and validation-based checkpoint selection. The ANFIS baseline (radius ????=0.5r=0.5, 19 rules) achieved RMSE/ MAE/ MAPE/ ????2 of 1191.63/ 751.02/ 0.001921/ 0.999996 on validation and 1207.23/ 664.70/ 0.001311/ 0.999996 on testing. Training required 546.91 s, while inference remained lightweight (0.0846 s for 100 validation samples and 0.1493 s for 300 test samples). Diagnostic analyses using learning curves, parity plots, residual inspection, and empirical error distributions further supported the strong agreement between predicted and observed RTT values. These results indicate that ANFIS with subtractive clustering can deliver accurate and low-latency RTT prediction suitable for QoS-aware orchestration pipelines where training can be performed offline.

The Effect of STEAM-Based Learning Assisted by Educational Games and Virtual Reality on Junior High School Students’ Learning Motivation in Basic Programming Materials

2026-04-20T06:35:27+00:00

This study is motivated by the need for technology-based learning innovations to enhance students’ learning motivation, particularly through the integration of the STEAM approach with educational games and virtual reality. This study aims to analyze the effect of STEAM-based learning assisted by educational games and virtual reality on students’ learning motivation. The method used was a quasi-experimental design with a pretest-posttest control group. The research subjects consisted of 56 students divided into an experimental class and a control class. Data were collected using a learning motivation questionnaire that had been tested for validity and reliability, while data analysis was conducted using multiple linear regression tests with IBM SPSS Statistics 27. The results showed that simultaneously, educational games and virtual reality had a significant effect (sig. 0.031 < 0.05) on students’ learning motivation. However, partially, educational games did not have a significant effect (sig. 0.091 > 0.05), and virtual reality also did not show a significant effect (sig. 0.984 > 0.05). The coefficient of determination (R²) value of 0.242 indicates that both variables contributed 24.2% to students’ learning motivation. These findings suggest that the integration of technology in STEAM learning has the potential to improve learning motivation; however, the effectiveness of each medium still needs to be optimized to produce a more significant partial effect.

Performance Evaluation of a Mobile Attendance System Using Dual-Factor Dynamic Qr Code and Gps Geofencing

2026-04-20T07:59:19+00:00

This study develops and evaluates an Android-based attendance system integrating dual-factor authentication using dynamic QR codes and GPS geofencing to prevent proxy attendance in higher education. The system was developed using the Waterfall model and evaluated through quantitative experiments measuring response time, GPS accuracy drift, and security robustness via Black-Box testing. Results show high efficiency with an average response time below 1.5 seconds. GPS validation achieved an average drift of 4.2 meters outdoors and 12.5 meters indoors, remaining within the 30-meter geofencing threshold. The system successfully rejected unauthorized attempts, including out-of-range scans and fake GPS spoofing. These findings demonstrate that combining dynamic QR codes with GPS validation significantly improves attendance authenticity and system reliability compared to single-factor methods. The study provides empirical evidence of a robust and scalable solution for secure mobile-based attendance systems in higher education.

A K-Prototypes-Based Approach for Modeling Student Segmentation Based on Learning Strategies to Support Academic Decision-Making

2026-04-20T08:04:02+00:00

This study aims to model student segmentation based on learning strategies using the K-Prototypes clustering algorithm. The data used consist of mixed-type variables, including categorical variables (gender and major) and numerical variables such as grade point average (GPA), learning habits, motivation, learning environment, health and social support, academic involvement, and academic achievement.

The analysis was conducted through several stages, including data preprocessing, exploratory data analysis, and clustering using the K-Prototypes algorithm. The optimal number of clusters was determined using the Elbow and Silhouette methods, both of which indicated that four clusters provide the best clustering structure.

The results show that students can be grouped into four distinct clusters with different characteristics. Cluster 3 represents highly motivated and high-achieving students with strong engagement, while Cluster 1 consists of students with good academic performance supported by favorable learning conditions. Cluster 4 includes students with moderate characteristics, and Cluster 2 represents students with lower performance and weaker learning strategies.

The clustering results were further validated using t-SNE visualization, which shows a reasonably clear distribution of clusters despite some overlap. Overall, this study demonstrates that the K-Prototypes algorithm is effective in handling mixed-type educational data and can provide meaningful insights to support data-driven academic decision-making and the development of targeted learning strategies.