Jurnal Politeknik Negeri Batam (PoliBatam)
Not a member yet
3001 research outputs found
Sort by
Comparison of KNN and Naïve Bayes Classification Algorithms for Predicting Stunting in Toddlers in Banjaran District
No full textStunting is a chronic nutritional problem that seriously impacts child growth and development. This study aims to compare the performance of the Naïve Bayes and K-Nearest Neighbors (KNN) algorithms in predicting stunting in toddlers in Banjaran District. The dataset consists of 12,000 toddler data points with three main features: age, gender, and height. The research employed a quantitative approach by applying machine learning algorithms. The SMOTE oversampling technique was applied only to the training data to avoid data leakage, and 5-fold cross-validation was used. A K-value of 3 was selected for the final KNN model based on validation curve analysis to prevent overfitting. The results show that KNN significantly outperformed Naïve Bayes across all evaluation metrics. The Naïve Bayes model yielded an accuracy of 67.50%, precision of 50.87%, recall of 61.38%, F1-score of 55.63%, specificity of 70.54%, and an AUC score of 75.71%. Meanwhile, the KNN (K=3) model achieved an accuracy of 99.11%, precision of 98.08%, recall of 99.25%, F1-score of 98.66%, specificity of 99.03%, and an AUC score of 99.65%. The performance difference between the two models was confirmed by McNemar\u27s Test with a p-value < 0.05, indicating a statistically significant difference. The low performance of Naïve Bayes was attributed to the violation of the feature independence assumption, particularly the high correlation between age and height (r ≈ 0.87). In conclusion, KNN is the more appropriate algorithm for stunting prediction on this dataset. However, the limitation of features suggests the need for further research with additional variables and external validation before wider-scale implementation.Stunting is a chronic nutritional problem that seriously impacts child growth and development. This study aims to compare the performance of the Naïve Bayes and K-Nearest Neighbors (KNN) algorithms in predicting stunting in toddlers in Banjaran District. The dataset consists of 12,000 toddler data points with three main features: age, gender, and height. The research employed a quantitative approach by applying machine learning algorithms. The SMOTE oversampling technique was applied only to the training data to avoid data leakage, and 5-fold cross-validation was used. A K-value of 3 was selected for the final KNN model based on validation curve analysis to prevent overfitting. The results show that KNN significantly outperformed Naïve Bayes across all evaluation metrics. The Naïve Bayes model yielded an accuracy of 67.50%, precision of 50.87%, recall of 61.38%, F1-score of 55.63%, specificity of 70.54%, and an AUC score of 75.71%. Meanwhile, the KNN (K=3) model achieved an accuracy of 99.11%, precision of 98.08%, recall of 99.25%, F1-score of 98.66%, specificity of 99.03%, and an AUC score of 99.65%. The performance difference between the two models was confirmed by McNemar\u27s Test with a p-value < 0.05, indicating a statistically significant difference. The low performance of Naïve Bayes was attributed to the violation of the feature independence assumption, particularly the high correlation between age and height (r ≈ 0.87). In conclusion, KNN is the more appropriate algorithm for stunting prediction on this dataset. However, the limitation of features suggests the need for further research with additional variables and external validation before wider-scale implementation
Comparative Analysis of Penetration Testing Frameworks: OWASP, PTES, and NIST SP 800-115 for Detecting Web Application Vulnerabilities
No full textWeb application security faces increasingly complex challenges as digital architectures evolve, necessitating the selection of appropriate and effective penetration testing methods. This study presents a comparative analysis of the OWASP Testing Guide, PTES, and NIST SP 800-115 frameworks in detecting web application vulnerabilities. Through experiments on DVWA and OWASP Juice Shop, the frameworks were evaluated based on detection speed, vulnerability count, and severity. The results highlight a clear trade-off: OWASP proved the most efficient (85 minutes average, 59 total vulnerabilities), making it ideal for rapid assessments. PTES demonstrated the most comprehensive technical depth (63 vulnerabilities, highest severity) but required the most time, while NIST SP 800-115 (49 vulnerabilities) excelled in compliance and risk management integration. The study recommends selecting OWASP for efficiency, PTES for deep technical audits, and NIST for regulatory alignment.Web application security faces increasingly complex challenges as digital architectures evolve, necessitating the selection of appropriate and effective penetration testing methods. This study presents a comparative analysis of the OWASP Testing Guide, PTES, and NIST SP 800-115 frameworks in detecting web application vulnerabilities. Through experiments on DVWA and OWASP Juice Shop, the frameworks were evaluated based on detection speed, vulnerability count, and severity. The results highlight a clear trade-off: OWASP proved the most efficient (85 minutes average, 59 total vulnerabilities), making it ideal for rapid assessments. PTES demonstrated the most comprehensive technical depth (63 vulnerabilities, highest severity) but required the most time, while NIST SP 800-115 (49 vulnerabilities) excelled in compliance and risk management integration. The study recommends selecting OWASP for efficiency, PTES for deep technical audits, and NIST for regulatory alignment
Sentiment-Based Knowledge Discovery pada Aplikasi iPusnas Menggunakan Metode Machine Learning dan Deep Learning
No full textiPusnas is a digital library application developed by the National Library of the Republic of Indonesia since 2016, with over 1.5 million users. Despite its potential to improve literacy, the application has only received a rating of 2.0. This study conducted sentiment analysis on 7.596 reviews obatained through web scraping using the Google Play Scraper Library. The data then underwent preprocessing steps including case folding, data cleaning, tokenization, stopword removal, and stemming. Reviews were automatically labeled based on the rating score, where scores of 1-3 were categorized as negative, with 5.174 entries, and scores 4-5 as positive, with 2.422 entries. The dataset was split in an 80:20 ratio, with 80% for training, and 20% for testing. The machine learning models tested were SVM, Random Forest, CNN, LSTM, and RNN. The evaluation metrics included accuracy, precision, recall, F1-score, and confusion matrix. CNN and LSTM achieved the highest accuracy (82%), Random Forest and CNN achieved the highest precision (81%), RNN the highest recall (79%) and LSTM the highest F1-score (79%). McNemar test showed a significant difference between Random Forest and CNN, Random Forest and LSTM, and between RNN and LSTM, while CNN and LSTM, as well as CNN and RNN, showed no significant difference.iPusnas is a digital library application developed by the National Library of the Republic of Indonesia since 2016, with over 1.5 million users. Despite its potential to improve literacy, the application has only received a rating of 2.0. This study conducted sentiment analysis on 7.596 reviews obatained through web scraping using the Google Play Scraper Library. The data then underwent preprocessing steps including case folding, data cleaning, tokenization, stopword removal, and stemming. Reviews were automatically labeled based on the rating score, where scores of 1-3 were categorized as negative, with 5.174 entries, and scores 4-5 as positive, with 2.422 entries. The dataset was split in an 80:20 ratio, with 80% for training, and 20% for testing. The machine learning models tested were SVM, Random Forest, CNN, LSTM, and RNN. The evaluation metrics included accuracy, precision, recall, F1-score, and confusion matrix. CNN and LSTM achieved the highest accuracy (82%), Random Forest and CNN achieved the highest precision (81%), RNN the highest recall (79%) and LSTM the highest F1-score (79%). McNemar test showed a significant difference between Random Forest and CNN, Random Forest and LSTM, and between RNN and LSTM, while CNN and LSTM, as well as CNN and RNN, showed no significant difference
BUILDING RESILIENT AND INCLUSIVE MSMES: SYNERGIZING GOVERNMENT SUPPORT AND DIGITAL FINANCE ADOPTION
No full textThis paper focuses on understanding the role of government support and digital financial adoption mediated by financial inclusion affecting the financial performance of Central Java’s MSMEs. This paper’s quantitative explanatory approach design is based on the Resource-Based Theory (RBT). Data were collected from 220 MSME actors through a simple random sampling method and analyzed using SEM-PLS by SmartPLS. The results indicate that government support, digital financial adoption, and financial inclusion significantly affect financial performance directly and indirectly. Theoretically, this study contributes to developing the RBT theory and practically becomes a reference for stakeholders in formulating strategies to improve the financial performance of MSMEs based on digital technology.Riset ini berfokus pada pemahaman mengenai peran government support dan digital financial adoption yang dimediasi oleh financial inclusion dalam mempengaruhi financial performance UMKM di Jawa Tengah. Desain pendekatan kuantitatif eksplanatori dalam riset ini berlandaskan pada Resource-Based Theory (RBT). Data dikumpulkan dari 220 pelaku UMKM melalui metode simple random sampling dan dianalisis menggunakan SEM-PLS dengan perangkat lunak SmartPLS. Hasil penelitian menunjukkan bahwa government support, digital financial adoption, dan financial inclusion secara signifikan mempengaruhi financial performance secara langsung dan tidak langsung. Secara teoritis, penelitian ini berkontribusi pada pengembangan konsep RBT dan secara praktis menjadi acuan bagi pemangku kepentingan dalam merumuskan strategi untuk meningkatkan financial performance UMKM berbasis teknologi digital
Analysis of Stacking Ensemble Method in Machine Learning Algorithms to Predict Student Depression
No full textMental health issues, particularly depression among university students, require early detection and intervention due to their profound impact on academic performance and overall well-being. Although machine learning has been utilized in previous studies to predict depression, most research still relies on single-model approaches and rarely employs publicly available datasets that have undergone comprehensive preprocessing. This study aims to develop a depression prediction model for university students using a two-level stacking ensemble technique with cross-validation stacking, integrating Random Forest, Gradient Boosting, and XGBoost as base learners, and Logistic Regression as the meta-learner. A public dataset from Kaggle was utilized, consisting of 502 student records and 10 multidimensional predictor variables. Data preprocessing included cleaning, feature encoding, and standardization. Model performance was evaluated using accuracy, precision, recall, F1-score, and ROC-AUC metrics. The proposed stacking ensemble model achieved excellent performance, with an accuracy of 98.02%, ROC-AUC of 99.8%, precision of 96%, recall of 100%, and an F1-score of 98% for the depression class. These results demonstrate that the stacking ensemble method is highly effective for early depression detection among university students and has strong potential for implementation as a decision-support tool in academic environments.Mental health issues, particularly depression among university students, require early detection and intervention due to their profound impact on academic performance and overall well-being. Although machine learning has been utilized in previous studies to predict depression, most research still relies on single-model approaches and rarely employs publicly available datasets that have undergone comprehensive preprocessing. This study aims to develop a depression prediction model for university students using a two-level stacking ensemble technique with cross-validation stacking, integrating Random Forest, Gradient Boosting, and XGBoost as base learners, and Logistic Regression as the meta-learner. A public dataset from Kaggle was utilized, consisting of 502 student records and 10 multidimensional predictor variables. Data preprocessing included cleaning, feature encoding, and standardization. Model performance was evaluated using accuracy, precision, recall, F1-score, and ROC-AUC metrics. The proposed stacking ensemble model achieved excellent performance, with an accuracy of 98.02%, ROC-AUC of 99.8%, precision of 96%, recall of 100%, and an F1-score of 98% for the depression class. These results demonstrate that the stacking ensemble method is highly effective for early depression detection among university students and has strong potential for implementation as a decision-support tool in academic environments
Application of Multinomial Naïve Bayes for Sentiment Classification on Bukalapak Reviews
No full textThis study investigates sentiment analysis on user reviews from Bukalapak, a major Indonesian e-commerce platform, using the Multinomial Naïve Bayes (MNB) classifier. The study focuses on tackling the challenge of data imbalance and the linguistic complexities of Indonesian, such as slang, affixes, and negation, which are common in user reviews. Data was collected through web scraping from Bukalapak\u27s app on the Google Play Store, resulting in a dataset of 19,999 reviews. A structured preprocessing pipeline was employed, including text normalization, tokenization, stopword removal, stemming, and term frequency-inverse document frequency (TF-IDF) weighting to prepare the data. The sentiment analysis results show that the model performs well in categorizing neutral reviews (accuracy 81%), but struggles with positive and negative sentiments due to data imbalance, leading to lower accuracy for these categories. The study highlights the effectiveness of Multinomial Naïve Bayes in large-scale sentiment analysis tasks in the e-commerce domain, particularly for platforms with large volumes of user-generated content. The study also introduces SMOTE (Synthetic Minority Over-sampling Technique) for handling data imbalance and k-fold cross-validation for model evaluation, significantly improving the model’s reliability. The research concludes that sentiment analysis can greatly benefit e-commerce platforms by improving customer service, informing product management decisions, and providing valuable insights for business strategies
The Inefficiency of Political Connections Becomes a Boomerang in the Challenge of Sustainable Growth
No full textThis study investigates how political connections, board educational background, and carbon emission disclosure affect the Sustainable Growth Rate (SGR) of infrastructure firms listed on the Indonesia Stock Exchange from 2019 to 2023. SGR reflects a firm’s capacity for long-term growth consistent with its financial stability and strategic balance. Using panel data regression with the Random Effects Model (REM), selected based on the Hausman test, this research operationalizes political connections through the presence of politically affiliated commissioners and carbon emissions through disclosure indices. The results indicate that political connections significantly hinder sustainable growth, suggesting that politically connected boards may intensify agency conflicts and weaken governance effectiveness. Meanwhile, board education and carbon emission disclosure show emerging yet statistically weak positive effects, implying that intellectual capacity and environmental accountability are not yet fully embedded in corporate sustainability strategies. Despite the limited sample size, robustness checks confirmed the consistency of the results. Overall, these findings highlight that sustainable growth in Indonesia’s infrastructure sector depends more on governance integrity and managerial competence rather than political privilege. This study contributes to the literature by providing empirical evidence of how agency conflicts arising from political affiliations can undermine long-term corporate sustainability
Peningkatan Partisipasi Masyarakat dalam Pelaporan Kerusakan Fasilitas Publik di Kota Batam Melalui Aplikasi BALAP-IN
No full textFasilitas public di Kota Batam masih menghadapi sejumlah kendala, seperti kerusakanjalan, taman, mapun lampu penerangan sering kali lambat mendapatkan perbaikan. Rendahnya keterlibatan masyarakat dalam memnyampaikan laporan turut memperburuk kondisi ini, terutama karena mekanisme pelaporan yang dianggap berbelit dan kurang jelas. Salah satu solusi yang dapat dilakukan adalah dengan memanfaatkan teknologi berupa aplikasi BALAP-IN (Batam Lapor-Infrastruktur), dimana aplikasi ini merupakan platform digital yang dirancang untuk membantu masyarakat untuk melaporkan kerusakan fasilitas public secara praktis, transparan, dan terdokumentasi dengan baik. Oleh sebab itu, aplikasi BALAP-IN perlu diperkenalkan kepada masyarakat agar dapat dimanfaatkan sehingga diharapkan tumbuh kesadaran dan partisipasi aktif warga untuk melaporkan kerusakan yang ditemui kepada pihak pengelola. Kegiatan pengabdian menggunakan pendekatan partisipatif dengan menerapkan model teknologi partisipatif. Metode yang digunakan melalui sosialisasi di media sosial dalam bentuk kampanye digital, Brainstorming dan diskusi dengan pihak berwenang, dan FGD. Hasil kegiatan menunjukkan bahwa target dan tujuan tercapai yaitu: (1) Dapat menumbuhkan kesadaran masyarakat mengenai pentingnya melaporkan kerusakan fasilitas publik, melalui partisipasi masyarakat membagikan pengalamannya yang mencapai 70,87%; (2) aplikasi BALAP-IN dikenal masyarakat secara luas dan berpotensi dalam peningkatan partisipasi masyarakat dalam pelaporan kerusakan infrastruktur publik; (3) Antusias masyarakat dalam menggunakan aplikasi BALAP-IN, mengindikasikan dapat meningkatkan jumlah laporan kerusakan fasilitas publik yang valid di kota Batam. Dengan adanya aplikasi ini, diharapkan pengawasan dan pemeliharaan jalan di Kota Batam dapat berjalan lebih responsive, transpran dan berkelanjutan. Fasilitas public di Kota Batam masih menghadapi sejumlah kendala, seperti kerusakanjalan, taman, mapun lampu penerangan sering kali lambat mendapatkan perbaikan. Rendahnya keterlibatan masyarakat dalam memnyampaikan laporan turut memperburuk kondisi ini, terutama karena mekanisme pelaporan yang dianggap berbelit dan kurang jelas. Salah satu solusi yang dapat dilakukan adalah dengan memanfaatkan teknologi berupa aplikasi BALAP-IN (Batam Lapor-Infrastruktur), dimana aplikasi ini merupakan platform digital yang dirancang untuk membantu masyarakat untuk melaporkan kerusakan fasilitas public secara praktis, transparan, dan terdokumentasi dengan baik. Oleh sebab itu, aplikasi BALAP-IN perlu diperkenalkan kepada masyarakat agar dapat dimanfaatkan sehingga diharapkan tumbuh kesadaran dan partisipasi aktif warga untuk melaporkan kerusakan yang ditemui kepada pihak pengelola. Kegiatan pengabdian menggunakan pendekatan partisipatif dengan menerapkan model teknologi partisipatif. Metode yang digunakan melalui sosialisasi di media sosial dalam bentuk kampanye digital, Brainstorming dan diskusi dengan pihak berwenang, dan FGD. Hasil kegiatan menunjukkan bahwa target dan tujuan tercapai yaitu: (1) Dapat menumbuhkan kesadaran masyarakat mengenai pentingnya melaporkan kerusakan fasilitas publik, melalui partisipasi masyarakat membagikan pengalamannya yang mencapai 70,87%; (2) aplikasi BALAP-IN dikenal masyarakat secara luas dan berpotensi dalam peningkatan partisipasi masyarakat dalam pelaporan kerusakan infrastruktur publik; (3) Antusias masyarakat dalam menggunakan aplikasi BALAP-IN, mengindikasikan dapat meningkatkan jumlah laporan kerusakan fasilitas publik yang valid di kota Batam. Dengan adanya aplikasi ini, diharapkan pengawasan dan pemeliharaan jalan di Kota Batam dapat berjalan lebih responsive, transpran dan berkelanjutan.  
Peningkatan Kapasitas UMKM Bengkalis melalui Pelatihan Penghitungan Harga Pokok Produksi dan Penjualan
No full textThe main problems faced by the majority of micro, small, and medium enterprises are a lack of knowledge of systematic cost recording and an inability to determine the correct selling price. This community service activity aims to improve the understanding and skills of micro, small, and medium enterprises (MSMEs) in calculating the cost of production and the cost of goods sold in accordance with applicable regulations. The implementation method for this community service activity consists of a pre-test, delivery of materials through lectures, interactive dialogue, simulation of COGS calculation, and a post-test. The participants in this activity were 25 micro, small, and medium enterprises in Bengkalis Regency. The results of the activity showed a significant increase in participants\u27 understanding, as reflected in the average pre-test score rising from 49.1 to 77.5 in the post-test. Participants understood the cost components that make up COGS, differentiated between direct and indirect costs, and applied COGS calculation methods for both trading and manufacturing businesses. Overall, this activity succeeded in increasing MSMEs\u27 capacity to record costs and determine selling prices more accurately and professionally. It is hoped that this increased understanding will contribute to the development of more efficient, transparent, and sustainable businesses.Permasalahan utama yang dihadapi mayoritas usaha mikro, kecil dan menengah adalah kurangnya pengetahuan mengenai pencatatan biaya secara sistematis serta ketidakmampuan dalam menentukan harga jual yang tepat. Kegiatan pengabdian kepada Masyarakat ini bertujuan untuk meningkatkan pemahaman dan keterampilan pelaku usaha mikro, kecil dan menengah dalam menghitung harga pokok produksi dan harga pokok penjualan berdasarkan ketentuan yang berlaku. Metode pelaksanaan kegiatan pengabdian ini terdiri dari pre-test, pemberian materi melalui ceramah, dialog interaktif, simulasi penghitungan HPP, dan post-test. Peserta kegiatan ini terdiri dari 25 pelaku pelaku usaha mikro, kecil dan menengah di Kabupaten Bengkalis. Hasil kegiatan menunjukkan adanya peningkatan signifikan dalam pemahaman peserta, yang tercermin dari kenaikan rata-rata nilai pre-test sebesar 49,1 menjadi 77,5 pada post-test. Peserta mampu memahami komponen biaya yang membentuk HPP, membedakan biaya langsung dan tidak langsung, serta menerapkan metode perhitungan HPP, baik untuk usaha dagang maupun manufaktur. Secara keseluruhan, kegiatan ini berhasil meningkatkan kapasitas pelaku UMKM dalam melakukan pencatatan biaya dan penetapan harga jual secara lebih tepat dan profesional. Diharapkan, peningkatan pemahaman ini dapat berkontribusi pada pengembangan usaha yang lebih efisien, transparan, dan berkelanjutan
PRODUCTIVITY IMPROVEMENT LAPPING POT PADA MESIN AUTO CONTOUR LAPPING
No full textImproving seal production output requires productivity enhancements that increase the efficiency and effectiveness of both machinery and tooling. One of the main issues was identified in the auto contour lapping machine, particularly in the lapping pot, which was considered ineffective and inefficient in its application. This condition resulted in low production output, longer processing times, and increased material waste due to the use of rubber-based lapping pots. To overcome these limitations, an improved lapping pot was designed and manufactured using 3D printing technology, replacing rubber with filament material to reduce production waste. This study focuses on the design, fabrication, and feasibility evaluation of the improved lapping pot through direct implementation in the production process. Performance evaluation included visual inspection, free height measurement, and surface quality inspection of the seals. The results show that the improved lapping pot is suitable for production use and significantly enhances productivity. By accommodating four seals in a single lapping pot, the processing time was reduced by a total of 75 seconds per cycle. Consequently, production output increased fourfold from 144 seals per hour to 576 seals per hour, while the seal quality remained consistent and met customer-required dimensions and specifications.Peningkatan hasil produksi seal memerlukan sebuah Improvement productivity yang berpengaruh dalam peningkatan efisiensi dan efektifitas suatu mesin maupun tooling menjadi lebih efektif dan dinamis. PT. X melakukan Improvement productivity pada salah satu proses pembuatan seal, proses yang dimaksud ialah proses mesin auto contour lapping. Improvement mesin auto contour lapping dilakukan pada bagian lapping pot yang kurang efektif dan efisien pada penggunaannya. Hal ini mempengaruhi jumlah produksi, waktu pengerjaan, dan menyebabkan adanya limbah saat melakukan pembuatan lapping pot. Untuk mengatasi masalah yang terjadi, peneliti membuat lapping pot yang lebih efektif dalam penggunaan. Pembuatan lapping pot yang sudah di Improve menggunakan mesin 3d printing dengan material filament yang membuat berkurangnya sampah sebab sebelumnya menggunakan material rubber. Didalam analisa peneliti berfokus pada pembuatan lapping pot dan hasil dari lapping pot tersebut layak atau tidak dalam penggunaan. Untuk mengetahui lapping pot maka akan diuji coba di dalam proses produksi. Setelah uji coba akan dilakukan pengecekan dengan cara pengecekan visual, pengecekan freheight dan pengecekan surface. Hasil dari lapping pot yang di Improve dapat digunakan dalam proses mesin, setelah dilakukan pembuatan lapping pot serta proses uji coba didapatkan waktu pengerjaan lebih singkat, Waktu awal lapping pot 1 pcs seal bisa memakan waktu sekitar 25 detik sedangkan setelah di Improve dengan penambahan 4 pcs seal lapping pot selama penelitian bisa memangkas waktu 100 detik. Untuk hasil dari proses seal pada lapping pot tidak jauh berbeda dengan seal lapping pot yang lama dan sudah masuk kedalam ukuran dan spec yang diinginkan customer