Workplace accidents in industrial environments continue to cause significant human, economic, and operational losses, making proactive hazard identification a critical priority for occupational health and safety (OHS) management. This study develops a workplace hazard-identification pipeline that automatically analyzes accident chronology texts using BERT (Bidirectional Encoder Representations from Transformers) as a contextual feature extractor and Random Forest as an ensemble classifier. The publicly available IHMStefanini Industrial Safety and Health dataset was used as the source corpus; after data-quality screening, stratified train-validation-test splitting (80:10:10), and class-imbalance handling using SMOTE applied only to BERT embedding vectors in the training partition, the final modeling matrix covered 14 Critical Risk hazard categories. Text preprocessing included controlled normalization, tokenization with bert-base-uncased, padding and truncation to 128 tokens, and contextual embedding extraction into 768-dimensional feature vectors. Experimental results on the held-out test set show that the proposed BERT-Random Forest model achieved an accuracy of 94.7%, precision of 93.8%, recall of 94.2%, and F1-score of 94.0%, outperforming TF-IDF with SVM, Word2Vec with LSTM, BERT with SVM, and standalone BERT fine-tuning baselines. Statistical comparison using the McNemar-Bowker paired error test confirmed that the performance difference between the proposed model and the strongest baseline was significant (p < 0.01). The main contribution of this study is not the generic superiority of a hybrid BERT-Random Forest architecture, but its practical adaptation for multi-class workplace hazard identification from industrial accident narratives with lower computational cost than full transformer fine-tuning. The proposed method can support automatic incident triage, hazard monitoring, prioritization of safety investigations, and decision.
Mitra program pengabdian Masyarakat ini adalah Masyarakat Desa Rejoso Lor. Dimana masalahnya adalah kurangnya pemanfaatan potensi lokal hasil budidaya ikan lele, kurangnya pemahaman manfaat dari olahan ikan lele dan kurangnya partisipasi Masyarakat dalam mengolah hasil potensi lokal yang ada di Desa tersebut. Tim dosen dari Universitas Merdeka Pasuruan dalam PKM memberikan sosialisasi, pelatihan dan pendampingan mulai dari pemahaman tentang manfaat dari olahan iken lele, kemudian olahan ikan lele dapat dijadikan dalam berbagai macam produk olahan sampai pada pembuatan kemasan dan pemasaran produk. Dari hasil PKM ini diharapkan Masyarakat dapat membuka banyak lapangan pekerjaan dan meningkatkan perekonomian Masyarakat desa.
This study analyzes developments in Value Added Tax (VAT) research in developing countries using a bibliometric approach and relates them to the Javanese philosophy of Hamemayu Hayuning Nagara, which emphasizes harmony between state interests and public welfare. Bibliometric data were obtained from the Scopus database for the 2014–2024 period, comprising 856 selected articles. The analysis was conducted through keyword mapping, author and country collaboration networks, co-citation analysis, document coupling, and thematic mapping using Biblioshiny in RStudio. The results show that the global VAT discourse is dominated by themes such as tax reform, value-added tax, and fiscal policy, with an increasing shift of focus toward issues of equity, redistribution, poverty, and inequality. The findings indicate that VAT is not merely a fiscal instrument, but also a tool of social and economic intervention used by governments to promote equity and achieve social justice. In the case of Indonesia, the increase of the VAT rate to 12% needs to be positioned within a framework of protecting vulnerable groups, ensuring fiscal transparency, and enhancing the effectiveness of social policies. Through the Javanese philosophy of Hamemayu Hayuning Nagara, an ideal VAT policy is understood as an effort to balance the need for state revenue with the moral obligation to safeguard public welfare. This study contributes by providing an intellectual map of VAT research and offering an ethical perspective based on local wisdom for the formulation of fair and sustainable tax policies
The increasingly fierce competition in the culinary business sector requires business actors to create a comfortable store atmosphere and provide high-quality services to increase customer satisfaction. This study aims to analyze the effect of store atmosphere and service quality on customer satisfaction at Mr. Blek Kolam Pancing dan Resto Pasuruan. This study uses a quantitative approach with an associative causal design. Data were collected using questionnaires distributed to 72 respondents selected via purposive sampling based on Hair's formula criteria. Data analysis techniques included instrument tests (validity and reliability), classic assumption tests, multiple linear regression analysis, coefficient of determination (R Square), F-test, and t-test. The results showed that store atmosphere and service quality simultaneously have a significant effect on customer satisfaction (F = 45,089, p < 0.05). Partially, both store atmosphere and service quality also significantly influence customer satisfaction. The coefficient of determination (R Square) of 0.657 indicates that 65.7% of customer satisfaction is influenced by store atmosphere and service quality, while the remaining 34.3% is influenced by other factors outside the study
Uncertainty regarding daily sales volume poses a challenge for inventory management in the culinary business, as it can lead to a mismatch between stock levels and demand. This study aims to apply the Seasonal Autoregressive Integrated Moving Average (SARIMA) method to forecast daily kebab sales using historical data from January 4, 2025, to January 3, 2026. The research process involved data preprocessing, splitting the data into training and testing sets, testing for stationarity using the Augmented Dickey-Fuller (ADF) test, identifying parameters via Autocorrelation Function (ACF) and Partial Autocorrelation Function (PACF) plots, and selecting the best model based on the Akaike Information Criterion (AIC) value. The ADF test results yielded a p-value of <0.05, indicating that the data was stationary. The optimal model identified was SARIMA (1,0,1)(1,0,1)₇, with an AIC value of 2509.27. Model evaluation resulted in an MAE of 15.09 portions, an RMSE of 17.09 portions, and a MAPE of 48.52%; these figures indicate that the average prediction error remains relatively high due to daily sales fluctuations. The model predicts sales of 26–28 portions per day, making it a useful reference for determining daily production volumes and raw material inventory requirements.
