cover
Article Per Year (5 Year)
All
Home Page
OAI Link
Editorial Team
Contact
Reviewer
Google Scholar
Contact Name
Ronal Watrianthos
Contact Email
ronal.watrianthos@gmail.com
Phone
+6281263621335
Journal Mail Official
joseitjournal@gmail.com
Editorial Address
Professional Organization - Ikatan Ahli Informatika Indonesia (IAII) / Indonesian Informatics Experts Association Jalan Jati Padang Raya No. 41 Jati Padang Pasar Minggu 12540 South Jakarta - Indonesia http://iaii.or.id/
Location
Unknown,
Unknown
INDONESIA
Journal of Systems Engineering and Information Technology
Published by Organisasi Profesi IKATAN AHLI INFORMATIKA INDONESIA
ISSN : -     EISSN : 2829310X     DOI : https://doi.org/10.29207/joseit.*
Core Subject : Science,
Computer Science & IT
International Journal of Systems Engineering and Information Technology (JOSEIT) is an international journal published by Ikatan Ahli Informatika Indonesia (IAII / Association of Indonesian Informatics Experts). The research article submitted to this online journal will be peer-reviewed. The accepted research articles will be available online (free download) following the journal peer-reviewing process. The language used in this journal is English. JOSEIT is a peer-reviewed, blinded journal dedicated to publishing quality research results in Computers Engineering and Information Technology but is not limited implicitly. All journal articles can be read online for free without a subscription because all journals are open-access.
Arjuna Subject : Ilmu Komputer (semua kategori) - Kecerdasan Buatan
Articles 4 Documents
 1 
Search results for , issue "vol. 3 no. 3 (2024)" : 4 Documents clear
Efficient Chain-of-Thought Guided Instruction Tuning for Graph Representation Learning Wang Yiming; LI Qian; Du Yuntao; Cui Lizhen; Yan Zhongmin
Journal of Systems Engineering and Information Technology (JOSEIT) Vol. 3 No. 3 (2024)
Publisher : Ikatan Ahli Informatika Indonesia Nusantara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29207/joseit.v3i3.8453

Abstract

Large language models (LLMs) reason well over natural language but do not naturally understand graph structure, and directly applying them to graph data runs into two persistent obstacles: a wide gap between the graph and text modalities, and the difficulty of encoding structural information in a way an LLM can use. Existing graph-instruction-tuning methods typically close this gap with simple linear projections and question-answer pairs that carry no explicit reasoning trace, which leaves cross-modal alignment shallow and drives up fine-tuning cost. This paper proposes GraphCoT, a graph representation learning method built on efficient chain-of-thought (CoT) fine-tuning. Its core is a high-quality CoT distillation mechanism in which a strong teacher model generates instruction data containing explicit intermediate reasoning paths, which explicitly teaches a student model to reason in multiple steps from only a small amount of training data. GraphCoT further introduces a graph-text alignment module that maps graph representations into the language embedding space, and a lightweight two-stage training strategy that aligns modalities before task-specific fine-tuning, which keeps training cost low while preserving both reasoning quality and transferability. Node classification and link prediction experiments on multiple benchmark datasets show that GraphCoT outperforms mainstream graph learning methods in accuracy and generalization, confirming the effectiveness of the CoT distillation mechanism and the alignment strategy for graph representation learning.
Quantifying Target Leakage in Rule-Derived Mental Health Screening Labels Rozi Meri; Ikhsan; Dian Eka Putra; Sofia Yosse; Ismael
Journal of Systems Engineering and Information Technology (JOSEIT) Vol. 3 No. 3 (2024)
Publisher : Ikatan Ahli Informatika Indonesia Nusantara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29207/joseit.v3i3.8470

Abstract

Publicly hosted tabular datasets increasingly package screening-questionnaire items alongside computed, not clinically observed, risk and severity labels. This creates a risk of target leakage: if a label is a deterministic function of the same features given to a classifier, reported predictive performance measures rule reconstruction, not screening validity. We audit a recent IEEE DataPort dataset (n = 2,005) scoring 27 Likert items across five conditions – ADHD, Autism Spectrum Disorder (ASD), Social Pragmatic Communication Disorder (SPCD), Depression, and Anxiety – each with a binary risk flag and a three-level severity label. The dataset's documentation discloses that every label is generated from a threshold or tertile rule on the summed items of its own condition, yet its usage instructions recommend fitting Random Forest or XGBoost with SHAP directly on the 27 raw items. We verify the documented rule empirically, then compare a trivial single-feature decision stump (the summed score) against a 300-tree Random Forest trained on all 27 items, under identical 5-fold stratified cross-validation, across all ten condition-by-task combinations (5 conditions × {risk, severity}). The single-feature stump matches or exceeds the 27-feature ensemble in every combination, reaching 100% cross-validated accuracy on nine of ten targets. Random Forest feature importance further shows 72–88% of predictive weight concentrated on each condition's own items. These results indicate the pipeline the dataset recommends reconstructs a disclosed arithmetic rule rather than learning a clinically meaningful pattern; any predictive-validity study on this dataset should exclude a condition's own summed items from that condition's feature set.
Predicting Student Exam Performance Using Explainable Regression Models: A Learning Analytics Study Thuy Thi Vo; Vu Thien
Journal of Systems Engineering and Information Technology (JOSEIT) Vol. 3 No. 3 (2024)
Publisher : Ikatan Ahli Informatika Indonesia Nusantara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29207/joseit.v3i3.8471

Abstract

Identifying which behavioral and demographic factors most strongly predict student academic performance is a long-standing goal of learning analytics. We apply linear regression, ridge regression, random forest, and gradient boosting to a 1,000-student behavioral and demographic dataset. Five-fold cross-validated R² shows that linear and ridge regression (R²=0.656) explain more variance than the ensemble methods (R²≈0.57), indicating a primarily additive structure in the data. To test this additive structure for residual non-linear or interaction effects and to obtain individual-level explanations, we apply SHapley Additive exPlanations (SHAP) to the random forest model via TreeExplainer, and cross-validate the resulting feature ranking against the standardized coefficients of the top-performing linear model. SHAP analysis identifies study time as the dominant predictor, with a mean |SHAP| of 4.93 points (approximately 4.1 exam-score points per additional study hour), followed by prior academic performance (mean |SHAP|=3.44) and attendance (mean |SHAP|=2.38). Internet access and part-time employment emerge as statistically meaningful binary factors (mean score gap: +5.0 and −3.3 points respectively), consistent with literature on the digital divide and academic workload. Parental education, gender, and extracurricular participation contribute negligible SHAP values (mean |SHAP| ≤0.16), a pattern that also holds for the corresponding linear coefficients. These results suggest that actionable behavioral variables – particularly study time and attendance – account for the bulk of explainable variance in examination outcomes, while fixed demographic characteristics contribute little incremental predictive power.
Near-Trivial Separability and Statistical Power Limitations in a Small Anthrax Symptom Dataset Khin Hnin Naing
Journal of Systems Engineering and Information Technology (JOSEIT) Vol. 3 No. 3 (2024)
Publisher : Ikatan Ahli Informatika Indonesia Nusantara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29207/6n7m2c61

Abstract

A recently released IEEE DataPort dataset frames 49 binary-labeled livestock symptom records as a resource for few-shot learning and robust classification under data scarcity. We audit the dataset's suitability for this role by computing per-symptom discriminative statistics and then assessing whether n=49 leaves meaningful room for the complex models the dataset is intended to evaluate. A single symptom – the presence of bloody discharge – reconstructs the anthrax label with 91.8% accuracy (sensitivity 77.8%, specificity 100.0%, MCC φ=0.83, 95% Wilson CI [80.8%, 96.8%]), and adding a second symptom (edema) raises this to 93.9%. A leave-one-out cross-validated logistic regression and a 300-tree random forest achieve 89.8% and 93.9% accuracy respectively, numerically no better than the trivial single-symptom rule; paired McNemar tests confirm that neither classifier's errors differ significantly from those of the trivial rule (p=1.00 for both). A power analysis shows that n=49 provides only 9.1% power to detect a 2.2-percentage-point improvement over the trivial rule at α=0.05. Reliably detecting even a 6.2-point improvement would require n≈90. These results indicate that this dataset cannot discriminate between trivial threshold rules and genuinely learned classifiers, and any reported accuracy near 92–94% on this dataset should be interpreted as evidence of near-trivial separability, not few-shot learning capability.

Page 1 of 1 | Total Record : 4

 1 

Filter by Year

2024 2024


Reset
Filter By Issues
All Issue Vol. 5 No. 1 (2026) Vol 4 No 2 (2025) Vol 4 No 1 (2025) Vol. 3 No. 3 (2024) Vol. 3 No. 2 (2024) Vol 3 No 2 (2024) Vol 3 No 1 (2024) Vol 2 No 2 (2023) Vol. 2 No. 1 (2023) Vol 2 No 1 (2023) Vol 1 No 2 (2022) Vol 1 No 1 (2022) More Issue