Universitas Ahmad Dahlan Journal
Not a member yet
5744 research outputs found
Sort by
Predictive Modelling for Mental Health Disorders using Machine Learning Techniques
Get PDFThis study evaluates the application of machine learning techniques in improving the prediction and diagnosis of mental health disorders. Traditional diagnostic methods are subjective and time-consuming, necessitating more accurate and efficient alternatives. Using a dataset from the Open-Sourcing Mental Illness survey, this study compares five machine learning algorithms-logistic regression, decision trees, random forests, k-nearest neighbours, and naïve bayes-on mental health prediction tasks. The findings indicate that Naïve Bayes achieves the highest accuracy (82.54%), suggesting its potential for more accurate mental health diagnostics. These results underscore the value of machine learning techniques in enhancing early detection and management of mental health conditions, paving the way for future research into more diverse datasets and ensemble approaches to refine predictive models for clinical application
Klasifikasi Kelayakan Keringanan UKT Menggunakan SMOTE dan Regresi Logistik
Get PDFKeringanan Uang Kuliah Tunggal (UKT) merupakan bantuan finansial bagi mahasiswa dari keluarga berpenghasilan rendah. Namun, proses seleksi penerima sering kali menghadapi tantangan subjektivitas dan ketidakseimbangan data, yang dapat berdampak pada ketepatan keputusan. Penelitian ini bertujuan membangun model klasifikasi untuk memprediksi kelayakan mahasiswa secara objektif menggunakan algoritma Regresi Logistik dan metode penyeimbangan data Synthetic Minority Over-sampling Technique (SMOTE). Penelitian ini menggunakan pendekatan kuantitatif dengan metode supervised learning. Dataset terdiri dari 100 data, yakni 80 data latih dan 20 data uji, dengan distribusi kelas yang tidak seimbang. Evaluasi model dilakukan menggunakan confusion matrix, akurasi, presisi, recall, dan F1-score. Hasil menunjukkan bahwa model tanpa SMOTE memiliki akurasi 91,2%, presisi 95,9%, recall 90,4% dan f1-score berada 93,1%. Setelah penerapan SMOTE, model menunjukkan akurasi meningkat menjadi 92,3% dengan presisi 94,0%, recall tetap stabil di nilai yang sama dengan model latih tanpa smote, dan F1-score mencapai 92,2%. Pada data uji, model mempertahankan kinerja tinggi dengan seluruh metrik evaluasi di atas 90%, menunjukkan kemampuan generalisasi yang baik dan minim overfitting. Penerapan SMOTE terbukti efektif dalam mengatasi ketidakseimbangan kelas dan meningkatkan sensitivitas model terhadap kelas minoritas
Analisis Perbandingan PCA-KNN dan SVM untuk Prediksi Risiko Diabetes
Get PDFDiabetes merupakan penyakit kronis yang sering terlambat terdiagnosis akibat gejala awal yang tidak spesifik, sehingga deteksi dini penting untuk mencegah komplikasi serius. Penelitian ini bertujuan menganalisis dan membandingkan performa kombinasi Principal Component Analysis dengan K-Nearest Neighbor (PCA-KNN) dan Support Vector Machine (SVM) dalam prediksi risiko diabetes. Dataset yang digunakan berasal dari Kaggle dengan 768 entri dan delapan atribut medis. Tahap praproses mencakup imputasi median untuk nilai nol, normalisasi Z-score, serta reduksi dimensi menggunakan PCA pada model KNN yang menghasilkan lima komponen utama dengan varian kumulatif >80%. Nilai k optimal ditentukan melalui 10-Fold Cross Validation dengan hasil terbaik pada k=16. Hasil evaluasi menunjukkan PCA-KNN mencapai akurasi 76,47%, sensitivitas 90,00%, dan spesifisitas 50,94%, lebih baik dibanding KNN standar. Sementara itu, SVM memperoleh akurasi 72,73% dengan spesifisitas tinggi (84,00%) namun sensitivitas rendah (51,85%). Temuan ini mengindikasikan bahwa PCA-KNN lebih sesuai untuk skrining awal karena sensitivitas tinggi, sedangkan SVM dapat digunakan pada tahap konfirmasi berkat spesifisitas yang lebih baik
Identification of Volatile Compounds in Lemon, Local Lemon and Lime Peel Extract Using Gas Chromatography – Mass Spectrometer
Get PDFOrange peel is one of the main sources of essential oil. Lemon orange can also be found in an area of Jambi City, commonly referred to as local lemon, but the shape and characteristics differ slightly from common lemons. The identification results of local lemon peels indicate that these oranges are a cross between lemons and limes (Citrus medica × Citrus aurantifolia). This study aims to analyze and determine the differences in the components of the peels of lemon, local lemon, and lime using gas chromatography–mass spectrometer (GC-MS). The contribution of this research lies in providing scientific data on the chemical composition of local lemon peels, which have not been widely studied, thereby offering potential applications for the development of natural products, essential oil industries, and local biodiversity utilization. This research was conducted by extracting the peels from lemons, local lemons, and limes using acetone as a solvent. The extracts were then analyzed for their components using GC-MS. The GC-MS analysis of acetone extracts from lemon, local lemon, and lime peels revealed 19 compounds in each sample. Four compounds were found to be common across all three samples, namely 2-pentanone, 4-hydroxy-4-methyl; β-bisabolene; bis (2-ethylhexyl) phthalate; and 2H-1-Benzopyran-2-one, 5,7-dimethoxy. Meanwhile, 15 other compounds showed different contents, indicating that the hybrid nature of local lemons influenced their chemical composition. These findings highlight the unique characteristics of local lemon peels and their potential as a valuable source of bioactive compounds
Development and implementation of kintung-based learning media in elementary music education
No full textThe limited availability of contextual and culture-based music learning media in elementary schools restricts students’ understanding of musical concepts and diminishes engagement with local cultural values. This study aims to develop and implement Kintung-based music learning media by utilizing the traditional bamboo musical instrument of the Banjar community to enhance students’ musical skills and cultural appreciation. The research adopted a Research and Development (R&D) design using the ADDIE model, with qualitative techniques embedded in each stage. During the Analysis phase, data were collected through interviews, observations, and documentation of Kintung practices in community and school settings involving musicians, art teachers, and elementary school sudents. The Design and Development stages involved selecting appropriate bamboo materials, cutting and shaping resonant tubes, tuning instruments to the diatonic scale, and preparing instructional materials aligned with learning objectives. The Implementation stage consisted of learning activities structured into preparation, presentation, practice, and performance, through which students learned instrument-playing techniques, rhythmic patterns, and traditional Banjar songs such as Ampar-Ampar Pisang and Ampat Si Ampat Lima. Evaluation results indicate that the media effectively improved students’ technical performance, rhythmic accuracy, ecological awareness, and appreciation of local culture. This study concludes that Kintung-based, music learning media provides a culturally grounded and pedagogically relevant innovation for integrating traditional music into elementary education, supporting students’ musical skill development, and contributing to the preservation of local cultural identity
Online gambling: Cross-border aspects and potential risk of divorce
Get PDFIntroduction to the Problem: Online gambling produces cascading social harms (debt, mental distress, and family conflict) that are surfacing in Indonesian divorce cases. Yet core enforcement gaps persist because gambling platforms, servers, and payment rails are frequently offshore and evidence is digital and volatile. Existing tools in the ITE Law and the Criminal Code lag behind these modalities.
Purpose/Study Objectives: To analyze how cross-border features of online gambling undermine Indonesian criminal and family-law responses, and to propose an integrated reform agenda that links criminal accountability with family protection.
Design/Methodology/Approach: Normative legal research combining statutory and conceptual analysis with comparative insights (licensed regimes such as Australia/UK; prohibition/ambiguous regimes) and illustrative Indonesian Religious Court decisions referencing gambling-driven marital breakdown.
Findings: Indonesia’s response is hampered by three enforcement deficits: (1) Platform/finance dependence: foreign digital platforms and domestic payment intermediaries (banks, e-wallets, telecoms) enable chip-based and crypto-denominated flows that current doctrine barely reaches; (2) Digital-evidence fragility: logs, metadata, and accounts are transient or hosted abroad, while preservation and admissibility standards and forensics capacity remain under-specified; and (3) Limited cross-border reach: narrow MLAT/extradition coverage and dual-criminality barriers where gambling is legal overseas. These deficits help explain a growing footprint of gambling in Indonesian divorce pleadings and judicial reasoning, even when causation is indirect (asset dissipation, coercive financial control, persistent conflict). Comparative practice shows courts can recognize gambling-related “wastage” in property division and maintenance, while regulators can harden payment and advertising controls. Overall, the paper finds that doctrinal silos between criminal/ITE rules and the Marriage Law weaken both enforcement and family protection.
Paper Type: Research Articl
Resolution of the Jiwasraya insurance case: Government perspective on ensuring legal certainty and justice
Get PDFIntroduction to the Problem: The Jiwasraya insurance scandal exposed major weaknesses in Indonesia’s legal oversight of state-owned enterprises, particularly in corporate governance, fiduciary responsibility, and regulatory enforcement. Despite multiple government interventions, the lack of accountability and transparency eroded public trust and questioned the integrity of legal policy.
Purpose/Objective of Study: This article examines the government’s legal and policy measures in addressing the Jiwasraya crisis, focusing on how these efforts align with the principles of legal certainty, justice, and Good Corporate Governance (GCG).
Design/Methodology/Approach: Employing a normative juridical method with statute and comparative approaches, the study analyzes statutory frameworks, court decisions, and administrative responses, supported by comparative insights from China, Germany, and the United Kingdom.
Findings: The findings reveal that government measures, such as corporate restructuring, the establishment of IFG Life, and criminal prosecution, remain largely reactive and lack structural reform. The study argues for the codification of fiduciary duties, strengthening corporate criminal liability, and the selective imposition of severe penalties in corruption cases causing extensive state losses. Furthermore, the absence of transitional legal norms and enforceable state guarantees leaves non-migrated policyholders without legal protection. These findings highlight the urgency of reforming Indonesia’s corporate and financial governance system to restore legal certainty and uphold justice.
Paper Type: Research Articl
Is Phubbing Gendered? Examining The Influence of Neuroticism and Gender in College Students
Get PDFThis study examined the relationship between neuroticism and phubbing among university students and explored the moderating role of gender. Guided by the stress and coping model and personality theory, a cross-sectional survey was conducted with 236 Indonesian students (53% female; 85.8% aged 20–23) using validated measures: the Generic Scale of Phubbing and the IPIP-BFM-25 emotional stability subscale. Descriptive analysis showed that phubbing scores were notably right-skewed, with a mean of 53.2 (SD = 17.94) and a higher median of 66.0. Neuroticism significantly predicted phubbing behavior (Estimate = 2.369, SE = 0.120, 95% CI [2.133, 2.605], p < .001), while gender showed no main effect (p = .120) and did not moderate the neuroticism–phubbing relationship (interaction p = .093). Independent t-tests and Mann–Whitney U tests confirmed no significant gender differences in either phubbing or neuroticism levels. These findings suggest that neuroticism is a more robust predictor of phubbing than gender, likely due to differences in emotional regulation. Given the use of convenience sampling and a cross-sectional design, the generalizability of findings is limited. Targeted interventions focusing on emotional coping may be effective in reducing phubbing, especially among individuals high in neuroticism
Mapping The Accessibility to Mental Health Providers in Low-Middle Income Countries: A Scoping Review
Get PDFThis scoping review outlines the availability of mental health providers in low-middle income countries (LMICs), highlighting disparities in provider distribution, obstacles to care, and innovative strategies to fill gaps in mental health services. A systematic search across ScienceDirect, EBSCOhost, and Scopus identified 3,213 articles using Boolean strings targeting mental health access, barriers, and LMICs. Six studies met the inclusion criteria (2020–2024, English, open-access) after screening via Rayyan software and PRISMA-Scr guidelines. Mental health provider density in LMICs remains critically low, ranging from 0.1 to 0.9 per 100,000 population across regions. Key findings include That Digital health integration has shown promise in India and Indonesia for managing schizophrenia and addressing pandemic-related burnout, but it has faced challenges such as limited internet access, low literacy, and device availability. School-based programs (e.g., Nepal’s HASHTAG initiative) demonstrated potential for adolescent mental health promotion through multicomponent, culturally adapted interventions. Rural engagement barriers in Mexico highlighted geographic isolation, poverty, and cultural stigma, with 82% of individuals lacking access to care despite need. Only 33.3% of Nigerian university students utilized mental health services due to cost concerns and confidentiality doubts. Addressing LMICs’ mental health crises requires context-specific strategies: scaling digital tools with offline capabilities, expanding school-based prevention programs, and improving rural service delivery through community-driven models. Policymakers must prioritize workforce training, infrastructure investment, and anti-stigma campaigns to bridge treatment gaps
Analisis Risiko Portofolio Optimal pada Saham Syariah Menggunakan Liquidity Adjusted Capital Asset Pricing Model (LCAPM) dengan Pendekatan Value at Risk Generalized Pareto Distribution (VaR-GPD): Studi Kasus: Saham Jakarta Islamic Index 70 (JII70) Periode Juni 2023 – Mei 2024
No full textPasar modal syariah di Indonesia mengalami pertumbuhan yang signifikan seiring meningkatnya kesadaran masyarakat terhadap investasi berbasis prinsip Islam. Meskipun sesuai dengan kaidah syariah, investasi di pasar modal tetap mengandung risiko yang perlu dianalisis secara menyeluruh, terutama terkait likuiditas dan fluktuasi harga. Penelitian ini bertujuan untuk menganalisis risiko portofolio optimal saham syariah menggunakan model Liquidity Adjusted Capital Asset Pricing Model (LCAPM) dan pendekatan Value at Risk berbasis Generalized Pareto Distribution (VaR-GPD). Objek penelitian meliputi saham-saham syariah dalam indeks Jakarta Islamic Index 70 (JII70) periode Juni 2023 hingga Mei 2024. Dari 22 saham yang konsisten, terpilih 13 saham dengan return positif untuk dianalisis. Hasil pemilihan berdasarkan rasio Return-to-Risk menghasilkan lima saham terbaik, yaitu TPIA, JPFA, AKRA, CTRA, dan EXCL, dengan proporsi portofolio berturut-turut sebesar 35,6%, 21,1%, 18,0%, 13,8%, dan 11,5%. Portofolio tersebut menghasilkan expected return harian sebesar 0,1970%, dan estimasi risiko menggunakan VaR-GPD menunjukkan potensi kerugian sebesar 2,8714% pada tingkat kepercayaan 95%. Hasil ini menunjukkan bahwa kombinasi LCAPM dan GPD memberikan gambaran risiko yang komprehensif bagi investor syariah.
Kata kunci: Saham Syariah, JII70, Portofolio Optimal, Liquidity Adjusted Capital Asset Pricing Model (LCAPM), Value at Risk-Generalized Pareto Distribution (VaR-GPD