| ..:: CALL ME::.. |
 |
Vol. 6 No. 3 (2026): September-November 2026
##plugins.themes.liteScholar.issue.totalArticles## | ##plugins.themes.liteScholar.issue.totalPages##
##plugins.themes.liteScholar.issue.coverage##
##plugins.themes.liteScholar.issue.foreign##
Indonesia (27)
Taiwan, Province of China (2)
United States (2)
China (1)
Malaysia (1)
Nigeria (1)
Published:
2026-09-14
Articles
-
The Prediction of Hospital Readmission in Diabetic Patients Using Random Forest, XGBoost, and Support Vector Machine with an Explainable AI Approach
Zulfaqar Zul, Ridho Amanda Putra, Hadnan Hardiansyah HardiansyahShow Abstract | Download PDF | DOI: 10.31004/jestm.v6i3.480 | Check in Google Scholar
Hospital readmission among patients with diabetes is an indicator of healthcare quality, treatment effectiveness, and hospitalization burden. Early prediction of readmission risk may support targeted interventions and reduce repeated hospital stays. This study developed prediction models using Random Forest, Extreme Gradient Boosting (XGBoost), and Support Vector Machine (SVM), with SHapley Additive exPlanations (SHAP) applied to interpret model predictions. The Diabetes 130-US Hospitals dataset comprised 101,766 medical records containing demographic characteristics, hospital visit history, laboratory results, diagnoses, and medication use. The outcome was defined as unplanned readmission within 30 days of discharge by recoding the original “<30,” “>30,” and “NO” categories. Data were divided into training and test sets at an 80:20 ratio using stratified sampling, preserving the class distribution of 11.16% readmitted and 88.84% not readmitted without resampling. Random Forest and XGBoost were trained on the full training set, whereas SVM used a stratified subsample of 10,000 training records because of computational constraints. All models were evaluated using the same test set. XGBoost achieved the best performance, with a balanced accuracy of 0.632 and ROC-AUC of 0.684. McNemar’s test indicated that its performance differed significantly from SVM but not from Random Forest. SHAP identified payer type, diagnosis categories, age group, prior inpatient history, and Clinical Risk Score as influential predictors. These findings demonstrate the feasibility of combining machine learning and explainable artificial intelligence for clinically interpretable readmission prediction. However, reliance on a single historical public dataset and the absence of external validation limit the model’s readiness for clinical deployment.
637-647 -
Analysis Of The Causes Of Suboptimal Steering Gear Performance Based On Maintenance Factors And The Quality Of Spare Parts On MV. Mulianim
Muhammad Dion Putra Widjaya, Ari Yudha Lusiandari, Sri Mulyanto Herlambang, Antonius Edy Kristiyono, Shofa Dai RobbiShow Abstract | Download PDF | DOI: 10.31004/jestm.v6i3.484 | Check in Google Scholar
The operational reliability of a ship's steering gear system is vital for navigation safety and vessel maneuverability, as regulated by International Maritime Standards (SOLAS). However, performance degradation often occurs due to inadequate maintenance practices and spare part constraints. This study aims to analyze the root causes of suboptimal steering gear performance on MV. Mulianim using a qualitative case study approach. Data were gathered through direct field observations, semi-structured interviews with chief and engine officers, and documentation audits (Engine Log Book and PMS records) over a 12-month period. Data validity was verified through source triangulation and analyzed using Miles and Huberman’s interactive model combined with Cause-and-Effect Analysis. The results revealed that steering gear hydraulic pressure dropped below the operational baseline (85–90 bar vs. standard 100 bar) due to high fluid turbidity, degraded viscosity, and internal bypasses. This condition was directly triggered by a 26.7% delay in Planned Maintenance System (PMS) tasks and severe spare part stockouts (25% critical spares unavailable). To prevent navigational hazards, strict adherence to PMS schedules and optimization of the spare part procurement pipeline are strongly recommended.
648-654. -
Website-Based Medicinal Plant Information System Design for Universitas Pahlawan Botanical Garden Using the Prototype Method
Cindy Fatika Sari, Safni Marwa, Hanantatur AdeswastotoShow Abstract | Download PDF | DOI: 10.31004/jestm.v6i3.478 | Check in Google Scholar
The Medicinal Plant Collection Garden of Universitas Pahlawan (UP) Botanical Garden manages approximately 60 species of medicinal plants as a means of conservation and education, yet its publication remains limited to physical banners and social media. The QR codes installed on-site also still direct visitors to the general Wikipedia platform, resulting in information that is less specific and whose scientific validity is difficult to verify. This study aims to design and build a Medicinal Plant Information System website as an independent and accurate information medium. The system was developed using the Prototype method to facilitate continuous interaction and feedback with users. The application was built using the Laravel 10 framework and a MySQL database, with three user roles: Visitor, Admin, and Person in Charge. Key features include a medicinal plant catalog, an activity photo gallery, news articles, public reviews, and a PDF report export module. Testing was conducted through Black Box Testing and User Acceptance Testing (UAT). The test results show that all feature functionalities ran successfully with a high level of user acceptance, indicating that this medicinal plant information system is feasible for implementation to optimize digital botanical literacy at the UP Botanical Garden.
655-668 -
Performance Analysis of OTR Machine In The Black Tea Leaf Rolling Process At PTPN IV Bah Butong Unit
Indra Prasta, Andriono ManaluShow Abstract | Download PDF | DOI: 10.31004/jestm.v6i3.488 | Check in Google Scholar
This study aims to analyze the performance of the Open Top Roller (OTR) machine in the black tea leaf rolling process at PTPN IV Bah Butong Unit. The OTR machine is essential in black tea processing, functioning to grind and roll tea leaves after the withering process to shape and release cell sap. The recorded variables were material input, scheduled batch duration, downtime, operating time, and good output. The batches processed 3,020 kg of withered leaves; recorded duration ranged from 40 to 50 min per batch, with a mean of 44.4 min, while total downtime was 5 min and good output was 3,010 kg. Availability, Performance, and Quality were 98.87%, 91.12%, and 99.67%, respectively, producing an OEE of 89.79%. Performance was the lowest component because most batches exceeded the selected ideal cycle time of 40 min, although all remained within the plant’s operational range of 40–50 min. Average loading was 302 kg per batch, equivalent to 80.53% of the 375 kg nominal maximum. These results identify cycle-time consistency, rather than downtime or recorded mass yield, as the immediate improvement priority. The result must nevertheless be interpreted as a short observational snapshot: the Quality factor represents mass yield rather than sensory or biochemical tea quality, and the widely cited 85% OEE benchmark is a comparative convention rather than a universal acceptance limit. The findings describe only this ten-batch observation window.
669-681 -
Advancements in Metal Organic Frameworks for Energy Storage Applications: A Review of Current Research and Future Directions
Md Samin Yeasar Zahan Sparsha, Usho Mohsin, Olakunle Ibrahim Oresegun, Reuben KabantiyokShow Abstract | Download PDF | DOI: 10.31004/jestm.v6i3.291 | Check in Google Scholar
Growing energy demand and environmental pressures require storage materials that combine reversible capacity, efficient transport, and durability. Metal–organic frameworks (MOFs) offer tunable metal nodes, organic linkers, and pore structures, enabling their use as active frameworks, composite components, and precursors to derived materials. This review examines their application in lithium-ion, sodium-ion, lithium–sulfur, and zinc batteries, supercapacitors, and hydrogen storage, together with computational and artificial intelligence approaches to material design. Representative results include 2305 mAh g−1 after 500 cycles at 0.1 A g−1 for a MOF-derived lithium-storage electrode and 97.9% capacity retention after 1000 cycles at 1 A g−1 for a SnPS3@C sodium-storage electrode. These benchmarks arise from different chemistries and test conditions and do not establish a universal performance ranking. The review relates reported improvements to conductive networks, accessible reaction sites, and accommodation of structural change, while examining low conductivity, instability, and limitations in performance normalization. Hydrogen adsorption is assessed by distinguishing material uptake from complete-system storage requirements. Computational screening is considered alongside synthetic accessibility, model uncertainty, and experimental validation. Future research should prioritize application-specific hybrid architectures, matched-condition comparisons, scalable preparation, and device-level assessment to establish which MOF design principles remain effective beyond laboratory demonstrations
682-698 -
Round-Off and Truncation Error Trade-Off in Finite-Difference Numerical Differentiation: Optimal Step-Size Determination
Ping-Yun Tung, Anggara Trisna Nugraha, Yu-Min LeeShow Abstract | Download PDF | DOI: 10.31004/jestm.v6i3.449 | Check in Google Scholar
Numerical differentiation by finite differences is an inherently ill-conditioned operation in which two antagonistic error sources, truncation and round-off, govern the attainable accuracy. The objective of this study is to quantify this trade-off and to determine the step size that minimises the total error of the centred finite-difference operator. Using the smooth test function f(x) = -0.1x4 - 0.15x3 - 0.5x2 - 0.25x + 1.2 evaluated at x0 = 0.5, the first derivative was approximated for step sizes from 100 to 10-10 in IEEE-754 double precision, and the true, truncation, round-off, and total errors were measured and compared with a Taylor-series error model. The results reproduce the characteristic U-shaped error curve: truncation dominates for large h with second-order slope, round-off dominates for small h with slope minus one, and the minimum true error of 5.42x10-12 occurs at h = 10-6, in close agreement with the analytical optimum hopt = (3ε|f|/M3)1/3. The centred scheme is shown to be markedly more accurate than the forward scheme. The study concludes that an a-priori optimal step size can be reliably predicted, providing practical guidance for derivative-based engineering computation.
707-713 -
Evaluation of The Drainage System at The University of Pahlawan Tuanku Tambusai Based on An Analysis of The Rainfall Catchment Area
Amelia Yusnita, Beny Setiawan, Aris FiatnoShow Abstract | Download PDF | DOI: 10.31004/jestm.v6i3.486 | Check in Google Scholar
Infrastructure development at the Universitas Pahlawan campus has led to land-use changes that potentially increase surface runoff and affect the performance of the drainage network. This study aims to analyze catchment area characteristics, calculate design runoff discharge, and evaluate the capacity of the existing drainage network by comparing design runoff discharge with channel capacity. A quantitative descriptive method was employed, utilizing hydrological and hydraulic analysis approaches. Data sources included 10 years of rainfall data from the Silam Station, topographic data and field observations, Rainfall analysis was conducted using probability distributions; runoff discharge was calculated using the Rational Method, and channel capacity was analyzed using Manning's Equation. The results indicate that the Universitas Pahlawan campus area is divided into four catchment zones (A1, A2, A3, and A4) with a total area of 4,27 hectares. Frequency analysis identified the Log Pearson III distribution as the best fit, yielding a design rainfall of 150.31 mm for a 10 year return period. Of the 12 channels evaluated, all channels were individually capable of accommodating the design runoff discharge; however, overtopping was identified in Channel 10, as this channel is located at the confluence of four flows from Zones A1 to A4 that converge at the same outlet point, causing the discharge from the four upstream channels to concentrate simultaneously and exceed the channel's holding capacity. This study provides information on the performance of the campus drainage system and serves as a basis for considering capacity improvements and regular channel maintenance.
734-745
