Garuda - Garba Rujukan Digital

A Grid-search Method Approach for Hyperparameter Evaluation and Optimization on Teachable Machine Accuracy: A Case Study of Sample Size Variation Malahina, Edwin Ariesto Umbu; Iriane, Gregorius Rinduh; Belutowe, Yohanes Suban; Katemba, Petrus; Asmara, Jimi
Journal of Applied Data Sciences Vol 5, No 3: SEPTEMBER 2024
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v5i3.290

This study aims to evaluate the effectiveness of the grid-search method in hyperparameter optimization on Teachable Machine (TM) using a varying number of image samples. The hyperparameters studied include epoch (e), batch size (b), and learning rate (l). A structured grid-search method approach will be applied to test 216 hyperparameter combinations across 6 categories of sample size per class, namely 10, 25, 50, 100, 250, and 500. The results showed that the optimal combination findings were obtained based on variations in the number of samples as follows: 10 samples using e:100, b:256, l:0.001 get an accuracy range of ≥ 90%; for 25 samples using e:500, b:16, l:0.001 get an accuracy range ≥ 97%; for 50 samples using e:100, b:512, l:0.001 get an accuracy range ≥ 88%; for 100 samples using e:500, b:32, l:0.001 get an accuracy range ≥ 88%; for 250 samples using e:50, b:16, l:0.001 get an accuracy range ≥ 92%, and finally 500 samples using e:500, b:256, l:0.001 get an accuracy range ≥ 96% and on average are able to achieve 100% accuracy from the detection test results of the best value performed for each sample variation of the image object. This research provides significant contributions or benefits in finding the optimal hyperparameter configuration, minimizing overfitting, and shortening the search time for TM accuracy in image classification, particularly in human face recognition. The findings support the development of more efficient and accurate TMs and provide practical guidance for finding better hyperparameter optimization using the grid-search method approach. The results of this study have implications for improving the effectiveness and accuracy of TM models and their development in mobile web applications

Implementasi Paralel Bubble Sort dengan Menggunakan MPICH2 dan Perbandingannya dengan Implementasi Sekuensial Rosani, Erna; Imaculata Inriani, Maria; Asmara, Jimi
Jurnal Informatika Vol 25 No 2 (2025): Jurnal Informatika
Publisher : Institut Informatika Dan Bisnis Darmajaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30873/jurnalinformatika.v25i17

This research is motivated by the need for efficient sorting algorithms for large amounts of data. The objective of this research is to implement the Bubble Sort algorithm in parallel using MPICH2 and compare its performance with sequential implementations. The research method involved testing parallel and sequential Bubble Sort programs on datasets of varying sizes (10, 100, 1,000, 10,000, and 100,000 elements). The experimental results show that parallel Bubble Sort performs better than sequential Bubble Sort, especially for large amounts of data. The speed-up achieved was 4.44 times faster on a dataset of 100,000 elements. However, for small data sets, sequential Bubble Sort is more efficient due to the communication overhead between processes in parallel Bubble Sort.

Breaking Connectivity Barriers: B-Smart as an Innovative Low-Bandwidth Mobile Learning Solution for Underserved Communities Katemba, Petrus; Sumarlin, Sumarlin; Asmara, Jimi; Naatonis, Remerta Noni
Journal Evaluation in Education (JEE) Vol 7 No 2 (2026): April
Publisher : Cahaya Ilmu Cendekia Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37251/jee.v7i2.2769

Purpose of the study: Access to digital learning resources remains a critical challenge in underserved communities, particularly those constrained by limited connectivity, inadequate infrastructure, and low-specification devices. This study aims to design, develop, and evaluate B-Smart, a low-bandwidth mobile learning application specifically engineered to bridge the digital learning gap in resource-limited educational environments. Methodology: A Design and Development Research (DDR) approach was employed, integrating an offline-first architecture, modular microlearning content, lightweight interface components, and data-efficient synchronization. Evaluation involved technical performance testing on low-end Android devices (1–2 GB RAM), usability testing using the System Usability Scale (SUS), pre-test and post-test learning assessments, and qualitative user feedback from 80 students and 15 teachers. Main Findings: B-Smart demonstrated reliable technical performance, with an average module loading time of 1.8 seconds, memory usage of 112 MB, weekly data consumption of 0.9–1.2 MB, and an offline access success rate of 98.7%. Usability evaluation yielded an SUS score of 82.4, while learning assessments revealed a mean post-test improvement of 24.6 points over pre-test scores, confirming significant knowledge gains across all user groups. Novelty/Originality of this study: These findings establish B-Smart as a novel, pedagogically sound, and technically efficient mobile learning solution tailored for low-bandwidth contexts. Unlike existing applications that depend on stable connectivity, B-Smart's offline-first, resource-efficient design ensures uninterrupted learning continuity in underserved regions. The study contributes a replicable development framework for scalable digital education initiatives, with practical implications for policymakers, educators, and developers seeking to advance equitable access to quality education in communities.

Breaking Connectivity Barriers: B-Smart as an Innovative Low-Bandwidth Mobile Learning Solution for Underserved Communities Katemba, Petrus; Sumarlin, Sumarlin; Asmara, Jimi; Naatonis, Remerta Noni
Journal Evaluation in Education (JEE) Vol 7 No 2 (2026): April
Publisher : Cahaya Ilmu Cendekia Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37251/jee.v7i2.2769

Purpose of the study: Access to digital learning resources remains a critical challenge in underserved communities, particularly those constrained by limited connectivity, inadequate infrastructure, and low-specification devices. This study aims to design, develop, and evaluate B-Smart, a low-bandwidth mobile learning application specifically engineered to bridge the digital learning gap in resource-limited educational environments. Methodology: A Design and Development Research (DDR) approach was employed, integrating an offline-first architecture, modular microlearning content, lightweight interface components, and data-efficient synchronization. Evaluation involved technical performance testing on low-end Android devices (1–2 GB RAM), usability testing using the System Usability Scale (SUS), pre-test and post-test learning assessments, and qualitative user feedback from 80 students and 15 teachers. Main Findings: B-Smart demonstrated reliable technical performance, with an average module loading time of 1.8 seconds, memory usage of 112 MB, weekly data consumption of 0.9–1.2 MB, and an offline access success rate of 98.7%. Usability evaluation yielded an SUS score of 82.4, while learning assessments revealed a mean post-test improvement of 24.6 points over pre-test scores, confirming significant knowledge gains across all user groups. Novelty/Originality of this study: These findings establish B-Smart as a novel, pedagogically sound, and technically efficient mobile learning solution tailored for low-bandwidth contexts. Unlike existing applications that depend on stable connectivity, B-Smart's offline-first, resource-efficient design ensures uninterrupted learning continuity in underserved regions. The study contributes a replicable development framework for scalable digital education initiatives, with practical implications for policymakers, educators, and developers seeking to advance equitable access to quality education in communities.

Title

Found 14 Documents
Search

Abstract

Abstract

Abstract

Abstract

Title Search

Found 14 Documents Search

Abstract

Abstract

Abstract

Abstract

Title

Found 14 Documents
Search