Publication Search

The rapid growth of e-commerce platforms in Indonesia, particularly Tokopedia, has resulted in a large volume of consumer reviews containing valuable information regarding customer perceptions and satisfaction. However, manual analysis of such reviews is inefficient and prone to subjectivity, necessitating an automated approach based on machine learning. This study aims to classify the sentiment of sports product reviews on Tokopedia into positive, negative, and neutral categories by applying Logistic Regression, Support Vector Machine (SVM), and Random Forest using the Term Frequency–Inverse Document Frequency (TF-IDF) approach. The data were collected through web scraping of Indonesian-language sports product reviews and processed through several preprocessing stages, including data cleaning, case folding, tokenization, stopword removal, and stemming. Feature representation was performed using TF-IDF to transform textual data into numerical vectors, after which the dataset was divided into training and testing sets with an 80:20 ratio. Model performance was evaluated using accuracy, precision, recall, and F1-score metrics. The results indicate that the application of TF-IDF significantly improves the performance of all models, with SVM consistently achieving the most optimal performance compared to Logistic Regression and Random Forest. These findings demonstrate that classical machine learning algorithms combined with TF-IDF remain highly effective for sentiment analysis of Indonesian-language text. The implications of this study are expected to assist sellers in understanding customer opinions, support consumers in making informed purchasing decisions, and serve as a foundation for the development of sentiment analysis and recommendation systems on e-commerce platforms.

Gestational Diabetes Mellitus (GDM) is a pregnancy-related metabolic disorder that poses health risks to both mother and fetus if not detected early, requiring accurate prediction methods for early screening and clinical decision-making. This study applies the Random Forest algorithm to detect GDM risk using clinical data from the Pima Indian Dataset. Data preprocessing included handling missing values, standardization, feature engineering, and a 70:30 train–test split. Two models were developed: a baseline and an optimized model using GridSearchCV hyperparameter tuning, validated with 5-fold cross-validation. Performance was assessed using a classification report, confusion matrix, and ROC–AUC. Results show that the optimized model outperforms the baseline, achieving 88% accuracy, an AUC of  93%, and average recall of 81%–85%. Compared to previous studies, this approach demonstrates improved predictive performance. The findings indicate that combining Random Forest with comprehensive preprocessing, feature engineering, and model optimization is effective and feasible for developing a medical decision support system for early GDM risk screening.

Diseases in primary health services exhibit complex spatial-temporal dynamics due to urbanization and population mobility. Conventional surveillance approaches are difficult to capture these patterns adaptively. Machine learning (ML) based on spatio-temporal modeling offers a solution with the ability to detect disease clusters automatically and with high precision. Research Objectives: This research aims to develop a machine learning model to detect disease hotspots from primary service data in Indonesia, with a focus on improving prediction accuracy, interpretability, and relevance of health policies. Methodology: The primary service dataset for 2024 (5,343 entries) was analyzed using three ML models Gradient Boosting Machine (GBM), Temporal Random Forest (TRF), and Multi-EigenSpot with spatial (village) and temporal (week, month) features. Performance evaluation includes predictive (AUC, F1-score) and spatial (Moran's I, Spatio-Temporal Correlation Index) metrics. Results: The results showed that Multi-EigenSpot achieved the best performance (AUC=0.91; F1=0.86), with the detection of dominant hotspots in Sungai Asam and Beringin Villages. Moran's I value of 0.63 indicates a strong spatial autocorrelation, while STCI=0.57 indicates moderate temporal stability. Conclusions: ML-based spatio-temporal models are effective in identifying hidden disease patterns and have the potential to be integrated into national digital surveillance systems. This approach supports precision public health by providing a scientific basis for real-time location- and time-based intervention policies.

This study aims to analyze the effect of welfare inequality on poverty levels on the island of Sumatra. Welfare inequality is measured using the Gini Index, while poverty levels are measured by the percentage of the poor population at the provincial level. This study uses a quantitative method with a panel data approach covering ten provinces on the island of Sumatra during the period 2020–2024. The analytical techniques used include panel data regression with fixed and random effects models, as well as classical assumption testing to ensure model validity. The results show that welfare inequality has a positive and significant effect on poverty levels, meaning that increasing inequality in income distribution tends to increase the number of poor people. This finding indicates that uneven economic growth can worsen the welfare of the community, especially low-income groups. Therefore, more inclusive and sustainable development policies are needed, particularly in increasing equitable access to education, health services, and productive employment opportunities to reduce inequality and poverty levels on the island of Sumatra.

Accurate wind turbine power curve modeling plays a crucial role in performance evaluation, energy yield estimation, and data-driven control strategies. However, actual power curves often exhibit non-linear behavior influenced by atmospheric variability, measurement noise, and SCADA anomalies, making conventional modeling approaches less effective. This study proposes an optimized logistic power curve model whose parameters are tuned using Particle Swarm Optimization (PSO) to improve predictive accuracy. The analysis uses the Wind Turbine SCADA Dataset from Kaggle, which undergoes extensive preprocessing including physical rule filtering, outlier detection with the Interquartile Range (IQR) method, anomaly removal, and smoothing of the power signal. A three-parameter logistic model is selected due to its ability to capture the typical S-shaped relationship between wind speed and power output. PSO is applied to identify optimal model parameters by minimizing the Mean Squared Error (MSE), utilizing 40 particles over 200 iterations. The optimized model achieves strong predictive performance with RMSE of 404.09, MAE of 179.96, and R² of 0.904 on the test set, indicating that more than 90% of the variability in actual power can be explained by wind speed. Residual analysis reveals heteroscedastic patterns and slight overestimation in mid-range wind speeds, yet overall model consistency remains high. Comparative evaluation against Linear Regression, Random Forest, and logistic modeling using curve_fit shows that the Logistic–PSO approach provides the most accurate and stable predictions. These findings demonstrate that combining logistic modeling with PSO offers an effective and robust method for data-driven wind turbine power curve optimization.

Information system security faces serious challenges due to increasingly complex cyber attacks. Intrusion Detection Systems (IDS) require efficient approaches to handle high-dimensional data such as the NSL-KDD dataset with 41 features. This study aims to implement the Genetic Algorithm (GA) for feature selection on the NSL-KDD dataset to improve the efficiency and accuracy of network attack detection. The method used is computational experimental research, involving data preprocessing, GA implementation for feature selection, building a classification model using Random Forest, and performance evaluation based on accuracy, precision, recall, F1-score, and computation time. The results show that GA successfully reduced features from 41 to 12 features (70.7% reduction), significantly improving computational efficiency. However, model accuracy slightly decreased from 0.4973 to 0.4951, indicating that while GA is effective for feature selection, the elimination of certain features may reduce classification capability. The implication of this study is that GA can be used as a tool to simplify intrusion detection models, but it should be combined with parameter optimization and data imbalance handling to achieve more optimal performance.  

Islamic economic instruments in the Indonesian economic system play an important role, such as waqf which can provide many benefits to society. Waqf instruments continue to innovate, where previously they were limited to land or buildings, now there are waqf instruments in the form of money. However, cash waqf is considered to be still not optimal because the potential of existing waqf is very far from the waqf that can be collected. This study aims to determine whether there is an influence of attitudes, subjective norms, performance expectations and social influences on intentions to endow money. The population of this study is the Muslim community in the city of Semarang. By taking a sample of 150 respondents by purposive non-random sampling. The data collection used a Googleform questionnaire, the data analysis method used SEM (Structural Equation Model) based on the SmartPLS 3.0 application. The results showed that the research model used was valid and reliable so that it was used properly. Based on the hypothesis test, the result is that attitude has a positive and significant direct effect on the intention of cash waqf. Subjective norms have a positive and not significant direct effect on cash waqf intentions. Performance Expectation has a positive and significant effect on cash waqf intentions. Social influence has a positive and insignificant effect on cash waqf intentions. The conclusion from this study is that the higher the attitude factor in a person which comes from beliefs in behaviour, the higher the intention of cash waqf, and the higher the performance expectation factor which comes from the level of confidence in obtaining profits in behaviour, the higher the intention of cash waqf. Likewise, subjective norm factors originating from social pressure have no influence on cash waqf intentions and social influence factors originating from social pressure in perceiving the use of the system have no influence on cash waqf intentions.

Maritime safety is a crucial aspect of the maritime industry, particularly in Indonesia as an archipelagic country. Although regulations regarding maritime safety are well-established, the frequency of maritime transport accidents in Indonesia remains high (KNKT, 2019–2023). This study aims to analyze and test the partial and simultaneous effects of the Port Authority’s Role (X1), Vessel Seaworthiness (X2), and International Safety Management (ISM) Code (X3) on Maritime Safety (Y) at Ketapang Port. The research method used is explanatory quantitative with multiple linear regression analysis. The sample involved 75 respondents, including ship crew members and Port Authority officers, selected using Simple Random Sampling. The analysis results indicate that the three independent variables have a positive and significant impact on Maritime Safety. The regression model shows an adjusted R Square value of 0.610. Among the three variables, the ISM Code (B = 0.410) is the most dominant factor, followed by the Port Authority’s Role (B = 0.389), and Vessel Seaworthiness (B = 0.186). These findings highlight that structured internal safety management (ISM Code) has the highest leverage, supported by strong regulatory oversight.

The rapid growth of open-source software (OSS) in machine learning (ML) has intensified the need for reliable, automated methods to assess project quality, particularly as OSS increasingly underpins critical applications in science, industry, and public infrastructure. This study evaluates the effectiveness of a diverse set of machine learning and deep learning (ML/DL) algorithms for classifying GitHub OSS ML projects as engineered or non-engineered using a SMOTE-enhanced and explainable modeling pipeline. The dataset used in this research includes both numerical and categorical attributes representing documentation, testing, architecture, community engagement, popularity, and repository activity. After handling missing values, standardizing numerical features, encoding categorical variables, and addressing the inherent class imbalance using the Synthetic Minority Oversampling Technique (SMOTE), seven different classifiers—K-Nearest Neighbors (KNN), Decision Tree (DT), Random Forest (RF), XGBoost (XGB), Logistic Regression (LR), Support Vector Machine (SVM), and a Deep Neural Network (DNN)—were trained and evaluated. Results show that LR (84%) and DNN (85%) outperform all other models, indicating that both linear and moderately deep non-linear architectures can effectively capture key quality indicators in OSS ML projects. Additional explainability analysis using SHAP reveals consistent feature importance across models, with documentation quality, unit testing practices, architectural clarity, and repository dynamics emerging as the strongest predictors. These findings demonstrate that automated, explainable ML/DL-based quality assessment is both feasible and effective, offering a practical pathway for improving OSS sustainability, guiding contributor decisions, and enhancing trust in ML-based systems that depend on open-source components.

This study aims to analyze the relationship between managerial ownership, institutional ownership, audit committee, and research and development (R&D) expenses on Intellectual Capital Disclosure (ICD) in healthcare sector companies listed on the Indonesia Stock Exchange (IDX) during the 2020–2024 period. Intellectual Capital Disclosure is essential as it reflects a company’s ability to manage knowledge, innovation, and human resources that serve as its competitive advantage. This research employs a quantitative approach using the total sampling method, where all healthcare sector companies that meet the criteria are included as samples. Secondary data were obtained from annual reports and analyzed using panel data regression with the assistance of Stata 19 software. Model selection was conducted through Chow, Hausman, and Lagrange Multiplier (LM) tests, with the results indicating that the Random Effect Model (REM) was the most appropriate model to use. The results show that managerial ownership, institutional ownership, and audit committee have negative and insignificant relationships with Intellectual Capital Disclosure. In contrast, research and development activities have a positive and significant relationship with Intellectual Capital Disclosure.

This study investigates the impact of Environmental, Social, and Governance (ESG) Risk Ratings on stock prices of companies listed in the ESG Leaders Index on the Indonesia Stock Exchange during the period 2020–2023. Using the Ohlson (1995) valuation model as the theoretical framework, the research examines the value relevance of financial information—proxied by Book Value per Share (BVPS) and Earnings per Share (EPS)—and non-financial information in the form of ESG risk ratings. The study employs purposive sampling, resulting in an unbalanced panel dataset of 120 firm-year observations. Panel regression analysis with the Random Effect Model (REM) is applied, supported by classical assumption tests and sensitivity analysis. The findings reveal that BVPS has a positive and significant effect on stock prices, highlighting its role as a stable and value-relevant measure for investors. By contrast, EPS shows a positive but insignificant relationship, confirming the declining relevance of earnings in the Indonesian market. Moreover, ESG Risk Ratings exhibit a negative but statistically insignificant effect, suggesting that while firms with higher ESG risks tend to be valued lower, sustainability considerations are not yet consistently incorporated into equity valuation by Indonesian investors. These results imply that financial fundamentals, particularly BVPS, remain the dominant factor in stock price determination, whereas ESG information has not yet achieved value relevance in the Indonesian context. The study underscores the need for stronger regulatory enforcement, standardized ESG disclosure, and greater investor awareness to enhance the integration of sustainability risks into capital market decision-making.

This study aims to: (1) determine the influence of teacher Self Efficacy on students’ learning motivation in Economics Social Studies; (2) determine the influence of teachers’ teaching creativity on students’ learning motivation; (3) determine the influence of teacher attitudes on students’ learning motivation; (4) analyze the simultaneous influence of teacher Self Efficacy, teaching creativity, and teacher attitudes on students’ learning motivation; and (5) identify the most dominant variable affecting students’ learning motivation at SMP Negeri 9 Marusu, Maros Regency. The research was conducted from September to November 2025 using an associative quantitative approach with an ex-post facto design. The sample consisted of 155 students selected through proportional random sampling. The research instrument was a Likert-scale questionnaire that had passed validity and reliability testing. Data were analyzed using classical assumption tests and multiple linear regression with the help of SPSS. The results of the study indicate that: (1) teacher Self Efficacy has a significant partial effect on students’ learning motivation (t = 3.606, Sig. = 0.000), but becomes insignificant in the simultaneous model (t = –0.531, Sig. = 0.596); (2) teaching creativity has a positive and significant partial effect (t = 5.852, Sig. = 0.000); (3) teacher attitudes have a positive and significant effect (t = 8.008, Sig. = 0.000) and are the most dominant variable, as shown by the highest regression coefficient (B = 1.456; Beta = 0.916); (4) simultaneously, teacher Self Efficacy, teaching creativity, and teacher attitudes significantly influence students’ learning motivation (F = 38.854, Sig. = 0.000); and (5) teacher attitudes are the most dominant predictor, as indicated by B = 1.456 and Beta = 0.916. This means that the more positive the teacher's attitude, the higher the students' learning motivation.

This research aims to analyze the significance of the influence of personality, courage to take risks, income expectations on students' interest in entrepreneurship at Slamet Riyadi University, Surakarta. The types of data used in this research are quantitative. The data sources used are primary and secondary data. The sample in this research was 100 students at Slamet Riyadi University, Surakarta. The method used in sampling was proportional random sampling. Data collection techniques in this research used questionnaires. Test the research instrument using validity and reliability tests. The classical assumption test uses tests: multicollinearity, autocorrelation, heteroscedasticity and normality. Data analysis techniques use multiple linear regression tests, t tests, F tests, and R2. The results of the validity and reliability tests show that all statements regarding personality, courage to take risks, income expectations and interest in entrepreneurship are declared valid because the p-value is < 0.05 and reliable because Cronbach's alpha is > 0.60. The results of the classical assumption test show that all variables have passed the multicollinearity, heteroscedasticity, autocorrelation and normality tests with normal distribution. The results of the regression analysis obtained the equation Y = .612 + 0.221 X1 + 0.326 X2 + 0.556 X3 + e. The results of the t test show that personality (X₁), courage to take risks (X₂), and income expectations (X₃) have a significant effect on students' interest in entrepreneurship at Slamet Riyadi University, Surakarta. The results of the F test showed that the regression model used in this research was correct. The R² test results show that the contribution of the independent variable to the dependent variable is 45.5%, the remaining 54.5% is influenced by other factors outside the variables studied  

The rapid growth of social media has yielded vast digital traces with high potential for improving corporate forensic auditing. Their utilization, however, lags behind through technological reliability, privacy, and adherence to the law. The aim of this study is to explore effective utilization of social media digital traces in forensic auditing and develop a functional framework that lags neither behind through technological efficiency nor adherence to the law and ethics. A mixed-method design was utilized, combining quantitative machine learning analysis with qualitative document analysis and semi-structured interview insight. Quantitative data drawn from social media digital traces were processed using Random Forest algorithm with SMOTE for class balancing, while qualitative data were processed using thematic analysis. The results indicated high model performance with 91.3% accuracy and AUC-ROC of 0.94, together with three emergent themes: digital integration, ethics and privacy, and regulation and legality. The results demonstrate that digital footprints may serve as an effective early and reliable indicator for fraud detection, provided they are accompanied by clear regulatory and ethical frameworks. Its principal contribution lies in the development of an operational model that combines machine learning with legal and ethical perspectives, a new strategy which matures methodological refinement and practical application in today's forensic auditing.

This study aims to analyze significant differences in learning motivation and critical thinking skills simultaneously between students who learn with differentiated learning and those who do not, determine the effective contribution of the implementation of differentiated learning on acid-base materials to learning motivation and critical thinking skills of students simultaneously, to determine the effective contribution of differentiated learning implementation on acid-base material to high school students' learning motivation, to determine the effective contribution of differentiated learning on acid-base material to high school students' critical thinking skills, to determine the level of achievement of students' critical thinking skills in experimental and control classes. This study used a quasi-experiment model with a pretest-posttest control group design. The sample of this study consisted of two classes, namely experimental class and control class, with random sampling. The experimental class used differentiated learning, while the control class used discovery learning. Critical thinking skills data were obtained from the description test data and learning motivation data through questionnaires. MANOVA test was used to analyze differences in critical thinking skills and student learning motivation in experimental and control classes simultaneously. Test of between subject effects is used to analyze differences in each dependent variable in experimental and control classes. Partial eta square test is used to analyze the effective contribution of differentiated learning to the dependent variable. Descriptive statistics to determine the level of achievement of critical thinking skills. The results showed there were significant differences in learning motivation and critical thinking skills simultaneously and individually between students who used differentiated learning and those who did not, the effective contribution of differentiated learning on acid-base materials to critical thinking skills and student learning motivation simultaneously was 16.9% (high), the effective contribution of differentiated learning on acid-base material to students' critical thinking skills is 6. 1% (medium), the effective contribution of differentiated learning on acid-base material to student learning motivation is 12% (medium), the level of achievement of critical thinking skills in the experimental class is in the good category, while in the control class in the sufficient category.

This study aims to analyze the influence of the role of technology, time management, providing incentives, and infrastructure on employee work productivity in the East Jakarta area. The data collection of this study was carried out by random sampling distributing questionnaires using the Likert scale with values from one to five choices, namely Strongly Disagree = 1, Disagree = 2, Neutral = 3, Agree = 4, and Strongly Agree = 5. 100 respondents who had criteria as employees in East Jakarta were obtained from the distribution of questionnaires. The data in this study was analyzed using a Partial Least Square (PLS)-based Equation Structural (SEM) Model with SmartPLS 4.0 software. Outer Model with calculations of Convergent Validity, Discriminant Validity, Composite Validity, Cronbach's Alpha and Inner Model with calculations of statistical T, R-Square, F-Square, and VIF. The results showed that technology has a significant positive influence on employee work productivity. In addition, time management has a positive but not significant effect on employee work productivity. The provision of incentives has a significant effect on increasing employee work productivity. And infrastructure also has a positive and significant effect on employee work productivity. These results provide useful insights for companies in understanding the factors that affect employee productivity.

The era of digital transformation requires office administration employees to master digital competencies and maintain high work motivation to achieve optimal productivity. This study analyzes the influence of digital competence and work motivation on the productivity of office administration employees in the Greater Bandung area. The main problem studied is the gap in understanding how digital competence and work motivation interact in influencing productivity, considering that the majority of previous studies examined the two variables separately. This study aims to analyze the partial and simultaneous influence of digital competence and work motivation on employee productivity. The research method uses a quantitative approach with an explanatory research design involving 420 respondents of office administration employees selected through proportionate stratified random sampling. The data collection instrument is a structured questionnaire with a 5-point Likert scale that has been validated using Confirmatory Factor Analysis and tested for reliability using Cronbach's Alpha. The data analysis technique uses multiple linear regression analysis by first conducting a classical assumption test to ensure the feasibility of the model. The results showed that digital competence had a significant positive effect on productivity with a regression coefficient of 0.398, work motivation had a significant positive effect with a coefficient of 0.425, and simultaneously the two variables explained 52.8% of the variation in employee productivity. These findings confirm that work motivation has a slightly more dominant influence than digital competence, indicating the importance of psychological factors in maintaining long-term productivity consistency. This research contributes to the development of human resource management theory by integrating Resource-Based View Theory and Self-Determination Theory, as well as providing practical implications for organizations to adopt a dual-track approach in employee development that combines digital competency training with a holistic motivation system.

This study examines the dynamics of Quality of Work Life, Workload, and Job Satisfaction in efforts to enhance Employee Productivity at PT. Bank Capital Indonesia, Tbk. Faced with increasingly complex industry challenges, companies need to effectively manage human resources to maintain competitiveness. Work-life balance, job satisfaction, and quality of work life are important factors that play a role in employee productivity, especially when workload increases. A quantitative approach with a descriptive correlational design was used in this study, with data collected through questionnaires and analysed using Structural Equation Modeling (SEM) with the AMOS software. The cluster random sampling technique was applied to select a sample proportionally. The findings show that Quality of Work Life and Job Satisfaction positively influence employee productivity, while Workload negatively impacts Job Satisfaction. However, Workload did not show a significant direct effect on Employee Productivity. Mediation analysis revealed that Job Satisfaction does not act as a mediator between Quality of Work Life and Employee Productivity. These findings suggest that improving Job Satisfaction, Quality of Work Life, and Work-life Balance can enhance employee productivity in the banking sector.

Regional Disaster Management Agency (RDMA) workers must be on standby and respond for 24 hours, especially if there are emergency. This causes demands on workers and workers become fatigue. The aim of this study is to analyze the factors that influence work fatigue in Tuban Regency’s RDMA workers. This research includes analytical and observational with a cross sectional design. The study was conducted on 56 workers with a simple random sampling system. The dependent variable studied was physical and mental fatigue, while the independent variables consisted of work factors (working hours and workload), emotional demands (responsibility), and organizational demands (work shifts). Data collection was carried out by direct measurement, questionnaires, interviews, documentation and observation then processed with Structural Equation Modeling-Partial Least Square (SEM-PLS). The factors that affect work fatigue are work factors (t statistics = 3.643 and p-values = 0.000) and organizational demands (t statistics = 3.086 and p-values = 0.002), while emotional demands have no effect (t statistics = 0.950 and p-values = 0.342). Loading factors of physical work fatigue (0.917) and mental work fatigue (0.916) are almost the same, so they have the same contribution as a measure of fatigue. Workers who experience work fatigue should conduct regular medical examinations, know the workload limit for one day, organize sufficient work time, regularly rotate work time, stretch in the middle of work activities, prepare work music and videos during breaks.

The daily exchange of informatics over the Internet has both eased the widespread proliferation of resources to ease accessibility, availability and interoperability of accompanying devices. In addition, the recent widespread proliferation of smartphones alongside other computing devices has continued to advance features such as miniaturization, portability, data access ease, mobility, and other merits. It has also birthed adversarial attacks targeted at network infrastructures and aimed at exploiting interconnected cum shared resources. These exploits seek to compromise an unsuspecting user device cum unit. Increased susceptibility and success rate of these attacks have been traced to user's personality traits and behaviours, which renders them repeatedly vulnerable to such exploits especially those rippled across spoofed websites as malicious contents. Our study posits a stacked, transfer learning approach that seeks to classify malicious contents as explored by adversaries over a spoofed, phishing websites. Our stacked approach explores 3-base classifiers namely Cultural Genetic Algorithm, Random Forest, and Korhonen Modular Neural Network – whose output is utilized as input for XGBoost meta-learner. A major challenge with learning scheme(s) is the flexibility with the selection of appropriate features for estimation, and the imbalanced nature of the explored dataset for which the target class often lags behind. Our study resolved dataset imbalance challenge using the SMOTE-Tomek mode; while, the selected predictors was resolved using the relief rank feature selection. Results shows that our hybrid yields F1 0.995, Accuracy 0.997, Recall 0.998, Precision 1.000, AUC-ROC 0.997, and Specificity 1.000 – to accurately classify all 2,764 cases of its held-out test dataset. Results affirm that it outperformed bench-mark ensembles. Result shows the proposed model explored UCI Phishing Website dataset, and effectively classified phishing (cues and lures) contents on websites.