Article Per Year (5 Year)
p-Index From 2021 - 2026
0.408
P-Index
This Author published in this journals
All Journal Jurnal Pendidikan Terapan Journal of Deep Learning, Computer Vision and Digital Image Processing
Ahmad Khairul Shiddiq
Universitas Negeri Makassar, Makassar, Indonesia
Humanities Computer Science & IT Education Social Sciences Other

Published : 2 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 2 Documents
Search

 1 
Short-Video Addiction and Academic Dishonesty Among University Students: The Mediating Role of Fear of Failure and Self-Regulation Factors Ahmad Khairul Shiddiq; Mustari S. Lamada; Dwi Rezky Anandari Sulaiman
Jurnal Pendidikan Terapan Vol 4, No 2 May (2026)
Publisher : Sakura Digital Nusantara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.61255/jupiter.v4i2.872

Abstract

Purpose: Academic dishonesty is a critical concern in higher education, increasingly exacerbated by digital distractions like short-video platforms. Recent evidence indicates 77% of students report fear of failure as a primary reason for cheating. Yet, the interplay between digital addiction, psychological self-regulation, and integrity remains underexplored in contemporary environments. Purpose: This study investigates relationships between short-video addiction, emotion regulation difficulties, time management, fear of failure, and academic dishonesty among university students. Methods: A quantitative cross-sectional approach involved 336 undergraduates from Universitas Negeri Makassar. Data were collected via online survey using validated scales and analyzed using Partial Least Squares Structural Equation Modeling (SEM-PLS) with SmartPLS 4 software. Findings: Short-video addiction significantly predicted academic dishonesty, with emotion regulation difficulties emerging as a significant indirect predictor. Fear of failure mediated these relationships, transforming psychological distress into unethical behavior. Unexpectedly, time management skills indirectly positively predicted dishonesty, contrary to hypotheses, suggesting an overcommitment trap among high-achievers facing performance pressure rather than protection against misconduct. Research Implications: Universities should prioritize reducing psychological stigma around failure and promoting digital wellbeing over punitive measures. Interventions targeting emotion coping may reduce misconduct. Educators should emphasize mastery learning over competitive grading. Additionally, policy reviews regarding student overcommitment are necessary. Conclusion: Short-video addiction directly drives unethical behavior, while emotion dysregulation operates indirectly through failure anxiety. Uniquely, time management skills indirectly correlate with increased dishonesty, indicating organization may serve grade preservation rather than learning integrity. Interventions must address psychological distress and digital wellbeing to foster honesty. Originality: This study uniquely integrates short-video addiction with self-regulation factors through fear of failure, uncovering a paradoxical time management role in Indonesian higher education contexts.
A Statistical and Deployment-Aware Evaluation of EfficientNet-Lite0 and MobileNetV3-Small for 114-Class Lontara Script Recognition Performance Ahmad Khairul Shiddiq; Ummul Khaeri Masna; Dwi Ihza Sari Akis; Sultan Budiluhur
Journal of Deep Learning, Computer Vision, and Digital Image Processing Volume 4 Issue 1 March 2026
Publisher : CV. Sakura Digital Nusantara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.61255/decoding.v4i1.1693

Abstract

Purpose: This study evaluates the accuracy, calibration, efficiency, and deployment trade-offs of EfficientNet-Lite0 and MobileNetV3-Small for 114-class Lontara script recognition.Methods: A public Kaggle dataset containing 11,483 images was divided into training, validation, and test subsets using a stratified 70:15:15 split. Both transfer-learning pipelines were assessed using accuracy, macro and weighted metrics, confusion matrices, Expected Calibration Error, confidence-based rejection, McNemar’s exact test, paired bootstrap confidence intervals, TensorFlow Lite conversion, post-training quantization, and preliminary CPU inference diagnostics. Findings: EfficientNet-Lite0 achieved 94.31% test accuracy and a macro F1-score of 0.9415, compared with 90.60% and 0.9023 for MobileNetV3-Small. The paired difference in correctness was statistically significant (p=1.5893 ×10^(-6)), and the bootstrapped accuracy difference was 3.71 percentage points with a 95% confidence interval of 2.21 to 5.22. EfficientNet-Lite0 showed a lower observed calibration error, although the between-model ECE difference was not conclusive. MobileNetV3-Small produced substantially smaller and faster float16 TensorFlow Lite models, while several visually similar classes remained vulnerable to systematic high-confidence errors.Research implications: Model selection should reflect whether recognition quality or computational efficiency is prioritized. Results are limited to one dataset split, one reported training run, and preliminary CPU diagnostics.Originality: The study integrates statistical comparison, calibration, selective prediction, quantization, deployment diagnostics, and class-level error analysis within one Lontara recognition experiment.
Co-Authors Dwi Ihza Sari Akis Dwi Rezky Anandari Sulaiman Mustari S. Lamada Sultan Budiluhur Ummul Khaeri Masna