Publication Search

This study aims to analyze the process and implications of changes in the General Budget Policy (KUA) and the Temporary Budget Priorities and Ceilings (PPAS) on the preparation of the Revised Regional Revenue and Expenditure Budget (APBD) for the Fiscal Year 2025 at the Regional Financial and Asset Management Agency (BPKAD) of Sikka Regency. The research employs a qualitative descriptive approach using secondary data in the form of planning and budgeting documents as well as internship activity results. The findings indicate that the preparation process of KUA–PPAS has been conducted in accordance with applicable regulations, starting from planning based on RPJMD and RKPD, formulation by the Regional Government Budget Team (TAPD), and discussions with the Regional House of Representatives (DPRD), supported by the SIPD system. Changes in KUA–PPAS are influenced by internal factors such as discrepancies in revenue and expenditure realization, program shifts, and the utilization of budget surplus (SiLPA), as well as external factors including central government policy changes, macroeconomic conditions, and emergency situations.

Diabetes Mellitus (DM) is a metabolic disease with an increasing prevalence and a risk of causing macrovascular complications such as stroke. This study aimed to describe the characteristics of Diabetes Mellitus patients with stroke complications based on CT-scan results at RSKD Dadi Makassar in 2024–2025. This was a descriptive observational study with a retrospective design using medical record data. Samples were taken using a total sampling technique, comprising 60 patients, and analyzed univariately using the Statistical Product and Service Solutions (SPSS) version 26. The results showed that the majority of respondents were in the late elderly group (>56 years) at 66.7%, female (51.7%), and from the Makassar ethnic group (63.3%). Most respondents experienced hyperglycemia in random blood glucose (66.7%) and had uncontrolled fasting blood glucose (68.3%) and HbA1c (76.7%) levels. CT-scan results were dominated by non-specific cerebral infarction (68.3%), with ischemic stroke as the most common type (93.3%), a length of stay of 5–10 days (53.3%), and right-sided hemiparesis as the most common clinical manifestation (45.0%). It was concluded that DM patients with stroke complications were generally elderly, female, had poor glycemic control, and were dominated by ischemic stroke with non-specific cerebral infarction. Abstrak. Diabetes Melitus (DM) merupakan penyakit metabolik dengan prevalensi yang terus meningkat dan berisiko menimbulkan komplikasi makrovaskular berupa stroke. Penelitian ini bertujuan untuk mengetahui gambaran karakteristik pasien Diabetes Melitus yang mengalami komplikasi stroke berdasarkan hasil CT-scan di RSKD Dadi Makassar tahun 2024–2025. Penelitian ini merupakan penelitian observasional deskriptif dengan desain retrospektif menggunakan data rekam medis. Sampel diambil dengan teknik total sampling sebanyak 60 pasien dan dianalisis secara univariat menggunakan Statistical Product and Service Solutions (SPSS) versi 26. Hasil penelitian menunjukkan mayoritas responden berusia lansia akhir (>56 tahun) sebanyak 66,7%, berjenis kelamin perempuan (51,7%), dan berasal dari Suku Makassar (63,3%). Sebagian besar responden mengalami hiperglikemia pada GDS (66,7%) serta memiliki kadar GDP (68,3%) dan HbA1c (76,7%) yang tidak terkontrol. Hasil CT-scan didominasi oleh infark serebri tidak spesifik (68,3%) dengan jenis stroke terbanyak berupa stroke iskemik (93,3%), lama rawat inap terbanyak 5–10 hari (53,3%), dan manifestasi klinis tersering berupa hemiparese dextra (45,0%). Disimpulkan bahwa pasien DM dengan komplikasi stroke umumnya berusia lanjut, berjenis kelamin perempuan, memiliki kontrol glikemik yang buruk, dan didominasi oleh stroke iskemik dengan gambaran infark serebri tidak spesifik.

Document-level sentiment analysis assigns a single polarity label to an entire review, often obscuring opinion diversity within multi-sentence submissions. This limitation is particularly evident in reviews of multi-service platforms, where users frequently express heterogeneous opinions toward different aspects of the platform in the same review. To address this challenge, this study proposes a sentence-level sentiment analysis framework for Indonesian Gojek app reviews collected from the Google Play Store. The proposed framework introduces a two-stage segmentation strategy that combines punctuation-aware rules with conjunction-aware splitting based on coordinating and adversative conjunctions (e.g., tapi [but], padahal [even though]) to identify opinion boundaries and decompose mixed-sentiment reviews into independently classifiable sentence units. A total of 14,730 raw reviews collected between May and July 2025 were subjected to data cleaning and quality filtering, resulting in 7,187 valid reviews that were further segmented into 14,187 sentence-level instances. Each instance was manually annotated by three annotators using a four-class labeling scheme consisting of app-positive, app-negative, app-neutral, and service categories. Sentiment-level inter-annotator agreement, computed on the subset of instances unanimously categorized as app-related by all three annotators (n = 4,384), achieved substantial agreement (Fleiss'  = 0.636). Hyperparameter optimization was conducted using Optuna with the Tree-structured Parzen Estimator (TPE) sampler across four experimental scenarios. The best performance was achieved by IndoBERTweet under Stratified K-Fold evaluation, attaining an accuracy of 0.751 and a macro F1-score of 0.729, outperforming all IndoBERT configurations. The results demonstrate the effectiveness of domain-adaptive pre-training on informal Indonesian text and highlight the value of conjunction-aware segmentation for preserving fine-grained opinion structures in mixed-sentiment reviews. These findings suggest that domain-aligned language representations provide a practical and effective solution for sentence-level sentiment analysis of Indonesian app reviews.

Breast cancer remains a major global health challenge, with treatment effectiveness often limited by drug resistance and the toxic side effects of chemotherapy on normal cells. The exploration of bioactive compounds from natural sources through a medicinal chemistry approach offers a promising alternative strategy. This study aims to examine the molecular mechanisms of action and Structure-Activity Relationships (SAR) of various natural compound scaffolds as potential breast anticancer agents. The method employed was a systematic narrative literature review of 15 recent scientific articles evaluating computational parameters, including molecular docking, as well as in vitro and in vivo activities. The results indicate that polyphenols, flavonoids such as quercetin and EGCG, and curcumin possess strong cytotoxic activity and high binding affinity toward cancer-related target macromolecules. SAR analysis demonstrates that key structural features, including the number and position of free phenolic hydroxyl groups, the presence of gallate ester groups, and conjugated diketone systems, play a crucial role in determining ligand receptor complex stability. These interactions are supported by hydrogen bonding, hydrophobic interactions, and favorable steric compatibility within receptor binding sites. Computational findings further suggest that structural optimization can enhance ligand selectivity and improve pharmacokinetic properties. This study concludes that natural phytochemical scaffolds have significant potential as lead compounds and provide a rational basis for Computer-Aided Drug Design in developing more potent, selective, multi-target, and safer breast anticancer therapies.

Poverty is a multidimensional problem that remains a major development challenge in Central Tapanuli Regency. Growth in Gross Regional Domestic Product (GRDP), increased realization of Foreign Direct Investment (PMA) and Domestic Direct Investment (PMDN), and zakat collection should be important instruments in reducing poverty rates. However, data from 2019–2023 shows a discrepancy, where the growth in these macroeconomic and social indicators is not accompanied by a decrease in the number of poor people. This study aims to analyze the influence of zakat, GRDP, PMA, and PMDN on poverty levels in Central Tapanuli Regency. The research method used is qualitative descriptive analysis with secondary data from BPS, BAZNAS, and BKPM as well as interviews with relevant parties. The results show that zakat does contribute to alleviating the burden on mustahik, but its role is still limited because the majority of distribution is consumptive. GRDP increased from Rp9.95 trillion in 2019 to Rp13.67 trillion in 2023, but the resulting economic growth is not yet inclusive. Foreign direct investment (FDI) tends to be oriented towards large capital with limited labor absorption, while domestic direct investment (PMDN) is closer to local needs but still less than optimal in poverty alleviation. Overall, the increase in zakat, GRDP, FDI, and PMDN has not been able to reduce the poverty rate, which actually increased from 376,474 people in 2019 to 489,760 people in 2023. Thus, a more inclusive development strategy, optimization of productive zakat, and investment policies that favor labor-intensive sectors and MSMEs are needed so that economic growth truly impacts poverty reduction.

Antimicrobial resistance is a persistent threat to hospital care, particularly when empirical therapy relies on broad-spectrum antibiotics without continuous evaluation of local use and susceptibility patterns. This literature review aimed to synthesize evidence from four Indonesian hospital-based studies regarding antibiotic utilization, rationality assessment, and bacterial resistance profiles. A structured narrative review was conducted using four selected articles provided by the author. Data were extracted for study design, setting, population, antibiotic evaluation method, dominant antibiotic classes, rationality indicators, resistance profile, and stewardship implications. The four studies were descriptive and hospital-based, using retrospective records, concurrent observation, ATC/DDD with DU 90%, qualitative rationality criteria, and antibiogram data. The synthesis showed a consistent concentration of antibiotic use in broad-spectrum groups, especially third-generation cephalosporins, penicillins, quinolones, and macrolides. In a Bandung public hospital, total antibiotic consumption reached 95,719.01 DDD, with penicillins, cephalosporins, quinolones, macrolides, and sulfonamides included in the DU 90% segment. In intensive care, ceftriaxone was the most frequently used antibiotic and most rationality indicators were appropriate, although clinically significant drug interactions were still identified. In pediatric acute respiratory infection inpatients, cefotaxime and ceftriaxone dominated empirical therapy. Resistance mapping in Denpasar highlighted relevant Gram-positive and Gram-negative pathogens and recommended antibiotics according to susceptibility levels. Overall, the reviewed evidence supports an integrated antimicrobial stewardship model combining ATC/DDD-DU 90% surveillance, qualitative rationality evaluation, antibiogram-based empirical guidance, and periodic feedback to prescribers.

Picky eating behavior is a common feeding problem among preschool-aged children. Children who exhibit picky eating tendencies often reject various types of food, especially new or unfamiliar ones, placing them at risk of an imbalanced intake of macronutrients such as energy, protein, fat, and carbohydrates. Inadequate nutritional intake over the long term may negatively impact a child's nutritional status and growth. This study aims to examine the relationship between picky eating behavior and the adequacy of macronutrient intake including energy, protein, fat, and carbohydrates and nutritional status among preschool children at Lab School 1 Kindergarten, State University of Surabaya. This research utilized a quantitative design with a cross-sectional approach. The study population consisted of 60 preschool children aged 4–6 years (48–73 months), selected using total sampling. After applying inclusion and exclusion criteria, a total of 37 respondents were included. Data were collected using the Child Eating Behavior Questionnaire (CEBQ) to assess picky eating behavior, interviews with the Semi-Quantitative Food Frequency Questionnaire (SQ-FFQ) to evaluate macronutrient intake, and anthropometric measurements to determine nutritional status. Data analysis was conducted using the Spearman Rank correlation test. The results showed a significant relationship between picky eating behavior and energy intake (p=0.002; r=0.495), fat intake (p=0.002; r=0.502), carbohydrate intake (p=0.006; r=0.443), and nutritional status (p=0.002; r=-0.493) among preschool children at Lab School 1 Unesa. However, no significant relationship was found between picky eating behavior and protein intake (p=0.064; r=0.307).

This study examines the legal protection of children's rights to inclusive education and its implications for the institutional governance of Early Childhood Education (ECE) in Indonesia. A normative juridical method with a descriptive-qualitative library-based approach is used to evaluate the coherence between macro-level child protection regulations and operational standards for school management. The data are entirely secondary, sourced from statutory laws, ministerial regulations, and pertinent scientific literature. The findings reveal a fundamental tension: the constitutional rights of children with special needs to access non-discriminatory ECE are robustly guaranteed by the 1945 Constitution, Law No. 35/2014 on Child Protection, and Law No. 8/2016 on Persons with Disabilities, yet a wide gap persists at the implementation level. This discrepancy arises because derivative ECE governance instruments including accreditation frameworks and curriculum standards still frame inclusion readiness as a voluntary component rather than a binding obligation. Consequently, ECE institutions encounter systemic barriers in human resource management, physical accessibility, and curricular flexibility. The study underscores the urgency of transitioning ECE management toward a Human Rights-Based Approach (HRBA) and recommends that the Ministry of Basic and Secondary Education reform accreditation instruments by embedding inclusive indicators as mandatory prerequisites for institutional feasibility, thereby aligning administrative governance with the fulfillment of children's constitutional rights.

The crisis of human spirituality, which separates the inner dimension from the universe, is at the root of the current global environmental crisis. The purpose of this study is to comprehensively examine how Sufi principles are applied in environmental conservation practices at the Shadana Islamic Boarding School in Padarincang, and to determine how the teachings of the book Fathul 'Arifin influence the formation of the students’ ecological ethics. This research was conducted qualitatively through a case study. Data were collected through document analysis, in-depth interviews, and participatory observation from September 22 to October 5, 2025. The results indicate that the concept of spiritual purification, or tazkiyatun nafs, is integrated with nature conservation by activating the seven lathaif. Environmental conservation, such as protecting springs and forests, is viewed as a form of spiritual devotion to the Creator, as human awareness of the elements of fire, water, wind, and earth fosters a strong inner connection with the macrocosm or the universe. The results indicate that the santri paradigm has shifted from anthropocentrism to theocentrism. This study proposes an “Eco-Sufism” model, grounded in the transformation of inner consciousness, as an alternative solution to address environmental degradation. This approach emphasizes that to foster sustainable harmony among humans, God, and the universe, ecosystem sustainability requires a strong foundation of spirituality.

The intake of macro and micro nutrients is very important because the intake of these macronutrients is the main contributor to energy which is the main source for muscle growth. The purpose of this study is to determine the intake of macro and micro nutrients in the Islamic Islamic Boarding School Padang Garugur Padang Lawas Utara. This type of research is a quantitative descriptive research with the aim of creating a description or descriptive of a state of the research object. Data collection was conducted through interviews using macro and micro nutrient tester questionnaires. The questionnaire was given to students and students. The data were analyzed univariate to determine the distribution and frequency. The results of the study showed that the intake of macronutrients of students was in the protein intake of 19 respondents in the poor category, carbohydrate intake as many as 7 respondents in the poor category, and fat intake in 10 respondents in the poor category. The intake of micronutrients, namely the intake of drinking water consumed by students, was 9 respondents in the category of lack. It is recommended that students need to consume a variety of food and beverage intake that has nutritional content, in order to meet nutritional needs.

The intake of macro and micro nutrients is very important because the intake of these macronutrients is the main contributor to energy which is the main source for muscle growth. The purpose of this study is to determine the intake of macro and micro nutrients in the Islamic Islamic Boarding School Padang Garugur Padang Lawas Utara. This type of research is a quantitative descriptive research with the aim of creating a description or descriptive of a state of the research object. Data collection was conducted through interviews using macro and micro nutrient tester questionnaires. The questionnaire was given to students and students. The data were analyzed univariate to determine the distribution and frequency. The results of the study showed that the intake of macronutrients of students was in the protein intake of 19 respondents in the poor category, carbohydrate intake as many as 7 respondents in the poor category, and fat intake in 10 respondents in the poor category. The intake of micronutrients, namely the intake of drinking water consumed by students, was 9 respondents in the category of lack. It is recommended that students need to consume a variety of food and beverage intake that has nutritional content, in order to meet nutritional needs.

Student dropout in higher education remains a persistent socioeconomic challenge, yet many predictive models reported in the literature are methodologically compromised by randomized cross-validation schemes that introduce temporal data leakage and artificially inflate predictive performance. This study proposes a longitudinal prescriptive learning analytics framework integrating three complementary methodological components: a Leave-One-Cohort-Out (LOCO) temporal validation protocol, a hybrid SMOTE-ENN class balancing strategy, and temporal velocity feature engineering derived from Learning Management System (LMS) behavioral trajectories. The framework was evaluated on a longitudinal dataset comprising 464,739 enrollment records and 77 features. Five predictive algorithms—XGBoost, LightGBM, CatBoost, Random Forest, and Logistic Regression—were comparatively assessed on a strictly isolated blind holdout cohort (2022), with CatBoost emerging as the champion estimator, achieving a PR-AUC of 0.8859, a Macro F1-Score of 0.9143, and the lowest Brier Score (0.0221), thereby demonstrating superior calibration and discriminative capability under severe class imbalance (93:7 ratio). Comprehensive ablation analysis revealed that temporal velocity features function not merely as additive predictors, but as a structural prerequisite enabling Synthetic Minority Oversampling Technique with Edited Nearest Neighbors (SMOTE-ENN) to generate high-quality synthetic boundary instances; removing these features reduced minority-class precision from 0.8302 to 0.6721. To operationalize predictive outputs into actionable intervention pathways, Diverse Counterfactual Explanations (DiCE) were implemented under a three-tier causal constraint architecture on 96 borderline high-risk students, generating 384 feasible intervention scenarios exclusively targeting forward-looking behavioral velocity metrics without constraint violations. Collectively, these findings advance the paradigm of prescriptive learning analytics by providing educational institutions with interpretable risk diagnostics and operationally feasible intervention guidance grounded in empirically validated behavioral and temporal dynamics.

Diabetes mellitus, as a persistent non-infectious metabolic pathology increasingly prominent in the global epidemiological spectrum, is fundamentally characterized by dysfunctional hyperglycemia rooted in primary disruptions of pancreatic insulin hormone synthesis or peripheral resistance to its biological action, thereby inducing dynamic imbalances in gluconeogenesis and glycolysis pathways; amid the hypothesis that body mass index (BMI) as a measure of central adiposity potentially mediates variability in serum glucose levels, this quantitative observational study with a cross-sectional design rigorously tests the causal relationship between BMI and fasting/random blood glucose concentrations in a cohort of 134 adult subjects with type 2 diabetes mellitus affiliated with primary care services at Puskesmas Bongo II, Boalemo Regency, through an inclusive total sampling recruitment strategy, precision anthropometric measurement instruments (height, weight, WHO BMI categorization), and laboratory-validated glucometric validation, with multivariate inferential processing based on the Chi-Square independence test at a Type I error rate of α=0.05 using the latest edition of the SPSS analytical suite; the demographic profile highlights female gender supremacy (90 individuals, 67.2%), the normoweight group (78 cases, 58.2%), concurrent with substantial glycemic elevation prevalence (78 subjects, 58.2%), but the crucial statistical output reveals a p-value of 0.831 (>0.05) that negates any probabilistically meaningful association, thus the substantive conclusion affirms the non-significance of the BMI-glucose relationship in this local context, while implying the dominance of alternative etiopathogenic factors such as hypercaloric macronutrient intake patterns, deficits in aerobic/anaerobic physical activity, non-adherence to multidisciplinary pharmacological protocols (e.g., metformin/oral hypoglycemics), and a comprehensive management paradigm integrating behavioral education, continuous monitoring, and personalized interventions to mitigate long-term cardiovascular risks across the diabetes mellitus spectrum.

Diabetes mellitus, as a persistent non-infectious metabolic pathology increasingly prominent in the global epidemiological spectrum, is fundamentally characterized by dysfunctional hyperglycemia rooted in primary disruptions of pancreatic insulin hormone synthesis or peripheral resistance to its biological action, thereby inducing dynamic imbalances in gluconeogenesis and glycolysis pathways; amid the hypothesis that body mass index (BMI) as a measure of central adiposity potentially mediates variability in serum glucose levels, this quantitative observational study with a cross-sectional design rigorously tests the causal relationship between BMI and fasting/random blood glucose concentrations in a cohort of 134 adult subjects with type 2 diabetes mellitus affiliated with primary care services at Puskesmas Bongo II, Boalemo Regency, through an inclusive total sampling recruitment strategy, precision anthropometric measurement instruments (height, weight, WHO BMI categorization), and laboratory-validated glucometric validation, with multivariate inferential processing based on the Chi-Square independence test at a Type I error rate of α=0.05 using the latest edition of the SPSS analytical suite; the demographic profile highlights female gender supremacy (90 individuals, 67.2%), the normoweight group (78 cases, 58.2%), concurrent with substantial glycemic elevation prevalence (78 subjects, 58.2%), but the crucial statistical output reveals a p-value of 0.831 (>0.05) that negates any probabilistically meaningful association, thus the substantive conclusion affirms the non-significance of the BMI-glucose relationship in this local context, while implying the dominance of alternative etiopathogenic factors such as hypercaloric macronutrient intake patterns, deficits in aerobic/anaerobic physical activity, non-adherence to multidisciplinary pharmacological protocols (e.g., metformin/oral hypoglycemics), and a comprehensive management paradigm integrating behavioral education, continuous monitoring, and personalized interventions to mitigate long-term cardiovascular risks across the diabetes mellitus spectrum.

This study examines the efficiency of financial ratios in assessing corporate performance across countries. Although financial ratios are widely used as concise indicators of profitability, liquidity, solvency, and market value, their interpretive accuracy may vary across institutional, regulatory, financial, and macroeconomic environments. The objective of this study is to conceptually evaluate whether financial ratios can function as universally comparable performance measures in heterogeneous cross-country settings. Using a qualitative literature-based method, this study synthesizes prior findings on financial ratio analysis, financial statement comparability, market efficiency, regulatory enforcement, and macroeconomic stability. The findings indicate that profitability, liquidity, solvency, and market-based ratios are context-dependent indicators rather than universally stable measures. Their efficiency is influenced by accounting standards, audit quality, leverage norms, tax systems, capital market maturity, and macroeconomic volatility. The study proposes a contextual framework for interpreting financial ratios according to their sensitivity to national conditions. The implication is that researchers, analysts, and investors should combine ratio analysis with institutional and macroeconomic diagnostics to reduce biased performance interpretation in cross-country corporate evaluation.

This study examines the determinants of the Jakarta Composite Index (JCI) based on three main macroeconomic factors namely inflation, the USD/IDR exchange rate, and the SBI interest rate (BI Rate) covering the period January 2020 to December 2025, in the context of post-COVID-19 pandemic recovery and global economic turmoil. A quantitative approach was employed using the Ordinary Least Squares (OLS) method, with 72 monthly observations derived from secondary data sourced from official institutions including Bank Indonesia (BI), the Central Statistics Agency (BPS), the Indonesia Stock Exchange (IDX), and the Financial Services Authority (OJK). Classical assumption tests were applied comprising the Jarque-Bera normality test, Variance Inflation Factor (VIF) for multicollinearity, Breusch-Godfrey for autocorrelation, White Test for heteroscedasticity, and Ramsey RESET for model specification. Partially, inflation, exchange rate, and BI Rate each demonstrate a positive and significant effect on the JCI (p < 0.05). Simultaneously, all three variables exert a significant combined influence on the JCI, with a coefficient of determination R² = 0.4414, indicating that the model explains 44.14% of the variation in the JCI. The remaining 55.86% is attributed to other variables outside the model. Classical assumption test results reveal violations of normality, autocorrelation, and heteroscedasticity assumptions, although the model is free from multicollinearity. These findings confirm that Bank Indonesia's monetary policy has a significant and measurable impact on capital market performance. Further research is recommended using more advanced time series models such as GARCH or VECM to address violations of classical assumptions and improve estimation efficiency.

Mental health issues among university students have become a growing concern, driven by academic pressures, career uncertainties, and complex social transitions. However, a large proportion of students remains reluctant to seek professional psychological support due to social stigma and limited access to institutional counseling services. This study proposes the design and implementation of an Android-based chatbot application that integrates mood tracking and sentiment analysis to continuously monitor the emotional states of university students. The system employs a fine-tuned RoBERTa (Robustly Optimized BERT Pretraining Approach) model trained on the EmoContext conversational dataset (SemEval-2019 Task 3), comprising 30,160 labeled three-turn dialogue instances across four emotion classes: angry, happy, sad, and others. The model was fine-tuned for four epochs using the AdamW optimizer with a learning rate of 2e-5 and a maximum sequence length of 128 tokens. Evaluation on a held-out validation set of 6,032 samples yielded an overall accuracy of 88.28%, a macro-average F1-score of 0.87, and a weighted-average F1-score of 0.88. Per-class F1-scores were 0.89 (angry), 0.83 (happy), 0.91 (others), and 0.86 (sad). The classified emotion is transmitted in real time to the chatbot response logic, which generates empathetic replies and personalized relaxation recommendations based on the detected mood. Primary data collection through questionnaires and interviews with 62 and 19 university students respectively confirmed the need for accessible digital mental health support. The results demonstrate that RoBERTa-based fine-tuning on conversational data provides a reliable foundation for real-time emotion-aware mental health chatbot systems.

Indonesian culinary comments on social media platforms such as Instagram are characterized by informal spelling, regional language mixing, slang expressions, and emojis, posing substantial challenges for automated sentiment classification. While IndoBERT has demonstrated strong performance across Indonesian natural language processing tasks, the contribution of individual preprocessing components to fine-tuning performance on informal text remains underexplored, particularly in the culinary domain. This study addresses this gap by conducting a systematic preprocessing ablation study on IndoBERT-Base fine-tuning for Indonesian culinary sentiment classification, accompanied by a comparative evaluation against Naive Bayes with TF-IDF, SVM with TF-IDF, and BiLSTM as representative baselines. A dataset of 3,500 manually labeled Instagram culinary comments across three sentiment classes was used, with a stratified 80/10/10 split. Six preprocessing variants were evaluated under identical experimental conditions to isolate the contribution of each component. The results show that slang normalization is the most impactful single preprocessing step, yielding a macro F1-score gain of +0.0609 over the no-preprocessing baseline, while the full pipeline achieves an accuracy of 0.8800 and a macro F1-score of 0.8465. IndoBERT-Base with the full pipeline outperforms all baselines across all evaluation metrics. Per-class analysis reveals that the negative class achieves the lowest F1-score of 0.7600, with sarcastic expressions and Banjar regional vocabulary identified as primary sources of misclassification. These findings indicate that preprocessing decisions have a measurable and non-uniform effect on IndoBERT fine-tuning performance. In this study, slang normalization provides the most substantial individual contribution in bridging the vocabulary gap between informal user-generated text and the model’s pre-training distribution.

Sentiment analysis for Indonesian regional languages faces two persistent challenges: labeled training data is extremely limited for most regional varieties, and transformer models pre-trained on Bahasa Indonesia do not generalize reliably to languages with substantially different morphological structures. Prior work on the NusaX benchmark has primarily relied on direct fine-tuning, treating each regional language independently and without exploiting linguistic proximity between related languages as a transfer signal. This paper proposes Language-Similarity-Guided Transfer (LSGT), a sequential fine-tuning strategy that first adapts a pre-trained model to a pivot language selected using character trigram similarity, followed by fine-tuning on the target language. Four transformer models are evaluated across all 12 NusaX languages using the official train/validation/test splits: IndoBERT, NusaBERT, mBERT, and XLM-R. Performance is evaluated using four metrics: accuracy, macro F1, macro precision, and macro recall. Experimental results show that LSGT improves macro F1 in 44 of 48 model-language combinations, demonstrating that the fine-tuning strategy itself is a major factor in low-resource cross-lingual sentiment classification. XLM-R benefits most strongly from LSGT, achieving an average improvement of +0.137 macro F1 and a peak gain of +0.298 on Madurese. SHAP-based token attribution analysis further reveals that predictions rely heavily on named entities and domain-specific nouns rather than sentiment-bearing vocabulary, indicating a dataset-level bias inherited from the original SmSA corpus and propagated through the NusaX translation pipeline.

The study aims to analyze the influence of financial distress, leverage, and macroeconomic fundamentals on financial reporting fraud in state-owned enterprises (SOEs) listed on the Indonesia Stock Exchange (IDX) for the 2020–2024 period. A quantitative approach coupled with multiple linear regression analysis was employed. A saturated sampling technique was used to select 23 companies with 115 observation units. The data used were secondary data from published financial reports on the IDX. The results indicate that financial distress and macroeconomic fundamentals have no effect on financial reporting fraud, while leverage has a positive effect on financial reporting fraud. The model in this study is able to explain 6.9% of the variation in financial reporting fraud, while the remaining amount is influenced by factors outside the model. These findings indicate that companies with high debt levels are more likely to commit financial reporting fraud, while companies with financial problems and high interest rates are less likely to commit financial reporting fraud.