Jurnal Politeknik Negeri Batam (PoliBatam)
Not a member yet
3001 research outputs found
Sort by
KINERJA BERKELANJUTAN UMKM DI BALI: PERAN LITERASI KEUANGAN INOVASI FINTECH DAN INKLUSI KEUANGAN
No full textMicro, Small, and Medium Enterprises (MSMEs) in the creative industry sector tend to be oriented towards short-term business decisions, which causes stagnation and lack of direction in long-term development. Lack of access to structured information on finance, market share, and business management is a major challenge for MSMEs. Bali is one of the provinces with a large number of MSMEs and has the potential to improve the Indonesian economy. This study aims to analyze the effect of financial literacy, fintech innovation, and financial inclusion on the sustainable performance of MSMEs in Bali. The research method uses a quantitative approach with financial literacy, fintech innovation, and financial inclusion as independent variables and sustainable performance of MSMEs as the dependent variable. Data were collected by distributing questionnaires to 100 MSMEs and analyzed using a structural equation model based on Partial Least Square (PLS). The results of this study are expected to provide insight for stakeholders in designing more effective policies to improve the competitiveness and sustainability of MSMEs through increasing financial literacy, utilizing financial technology, and expanding access to formal financial services.Usaha Mikro, Kecil, dan Menengah (UMKM) di sektor industri kreatif cenderung berorientasi pada keputusan bisnis jangka pendek, yang menyebabkan stagnasi dan kurangnya arah dalam pengembangan jangka panjang. Kurangnya akses terhadap informasi yang terstruktur mengenai keuangan, pangsa pasar, dan manajemen bisnis menjadi tantangan utama bagi UMKM. Bali merupakan salah satu provinsi dengan jumlah UMKM yang cukup banyak dan memiliki potensi untuk meningkatkan perekonomian Indonesia. Penelitian ini bertujuan untuk menganalisis pengaruh literasi keuangan, inovasi fintech, dan inklusi keuangan terhadap kinerja berkelanjutan UMKM di Bali. Metode penelitian menggunakan pendekatan kuantitatif dengan literasi keuangan, inovasi fintech, dan inklusi keuangan sebagai variabel independen serta kinerja berkelanjutan UMKM sebagai variabel dependen. Data dikumpulkan melalui penyebaran kuesioner kepada 100 UMKM dan dianalisis menggunakan model persamaan struktural berbasis Partial Least Square (PLS). Hasil penelitian ini diharapkan dapat memberikan wawasan bagi pemangku kepentingan dalam merancang kebijakan yang lebih efektif guna meningkatkan daya saing dan keberlanjutan UMKM melalui peningkatan literasi keuangan, pemanfaatan teknologi keuangan, serta perluasan akses terhadap layanan keuangan formal
Visualisasi 3D Exterior Masjid Penyengat
Get PDFThe Sultan Riau Grand Mosque, also known as the Penyengat Mosque, is an important landmark due to its historical architectural details and symbol of the grandeur of the Islamic kingdom in Riau-Lingga. The grand imam of the mosque stated that, although it can be used as a religious and historical tourist destination. Therefore, the author plans to create a 3D visualisation of the mosque\u27s exterior by employing the multimedia development life cycle (MDLC) method and conducting qualitative analysis with experts to ensure that the visualisation closely resembles the original. The process of creating the 3D visualisation of the Penyengat Mosque focused on important parts, such as the dome, minarets, and distinctive mosque decorations. The results of the study based on qualitative analysis indicate that the 3D visualisation of the exterior of the Penyengat Mosque is an accurate representation of the original structure, including architectural details and realistic textures that reflect the original building, although there are minor differences in the floor area.Masjid Agung Sultan Riau, yang juga dikenal sebagai Masjid Penyengat, merupakan landmark penting karena detail arsitektur historisnya dan simbol kemegahan kerajaan Islam di Riau-Lingga. Imam besar masjid tersebut menyatakan bahwa, meskipun dapat digunakan sebagai destinasi wisata religi dan historis. Oleh karena itu, penulis berencana untuk membuat visualisasi 3D eksterior masjid dengan menggunakan metode siklus pengembangan multimedia (MDLC) dan melakukan analisis kualitatif bersama para ahli untuk memastikan visualisasi tersebut mirip dengan aslinya. Proses pembuatan visualisasi 3D Masjid Penyengat berfokus pada bagian-bagian penting, seperti kubah, menara, dan hiasan masjid yang khas. Hasil penelitian berdasarkan analisis kualitatif menunjukkan bahwa visualisasi 3D eksterior Masjid Penyengat merupakan representasi akurat dari struktur asli, termasuk detail arsitektur dan tekstur realistis yang mencerminkan bangunan asli, meskipun terdapat perbedaan minor pada area lantai
Animasi Aset Permainan “Safe Space” Sebagai Edukasi Pencegahan Kekerasan Terhadap Anak
Get PDFViolence against children remains a serious issue that continues to occur in many environments. Unfortunately, there is still a lack of effective visual educational media that help children recognize signs of danger or unsafe situations. Children need learning tools that are engaging, easy to understand, and appropriate for their developmental stage. Therefore, this study focuses on the design and development of two-dimensional animation assets for an educational game titled “Safe Space” as a medium for educating children about violence prevention. This research employs a product creation method, emphasizing the visual design and 2D animation production processes that support the game\u27s content. The stages in this method include observation, concept formulation, character and environment design, and animation production. The process is divided into pre-production, production, and post-production phases, which involve client brainstorming in pre-production, creation of 2D vector assets and animation using the cut-out technique during production, and rendering in post-production for game implementation. The final result of this project includes animated movements such as idle, jump, run, walk, slide, and fall, rendered as image sequences to be used as assets in the “Safe Space” educational game. These animations were created using the cut-out technique with the aid of Adobe Illustrator and Adobe After Effects. It is expected that the resulting animation will serve as an interactive and informative educational tool, helping children better understand preventive actions against violence.Kekerasan terhadap anak merupakan isu serius yang masih sering terjadi di lingkungan sekitar. Sayangnya, media edukatif visual yang secara efektif mengajarkan anak untuk mengenali tanda-tanda bahaya atau situasi yang tidak aman masih terbatas. Anak-anak membutuhkan media pembelajaran yang menarik, mudah dipahami, dan sesuai dengan perkembangan usia mereka. Oleh karena itu, penelitian ini berfokus pada perancangan dan pembuatan aset animasi dua dimensi untuk permainan edukatif berjudul “Safe Space” sebagai media edukasi pencegahan kekerasan terhadap anak. Penelitian ini menggunakan metode penciptaan produk, dengan menitikberatkan pada proses perancangan visual dan produksi animasi 2D yang mendukung konten permainan. Tahapan dalam metode ini meliputi observasi, perumusan konsep, desain karakter dan lingkungan, hingga proses animasi. Metode ini terdiri dari tahap pra produksi, produksi, dan pasca produksi, dimana tahap pasca produksi melakukan brainstorming Bersama klien, produksi membuat aset 2D vektor dan penganimasian dengan teknik cutout animation, serta pasca produksi tahap rendering yang siap digunakan untuk permainan. Hasil akhir dari penelitian ini berupa gerakan animasi idle, jump, run, walk, slide, dan fall yang dirender image sequence untuk digunakan sebagai aset permainan edukatif “Safe Space” yang dibuat menggunakan teknik cut-out dengan bantuan perangkat lunak Adobe Illustrator dan Adobe After Effects. Animasi ini diharapkan mampu menjadi media edukasi yang interaktif dan informatif bagi anak-anak dalam memahami tindakan preventif terhadap kekerasan
Comparative Study of Support Vector Regression and Long Short-Term Memory for Stock Price Prediction
Get PDFThis study aims to compare the performance of two machine learning algorithms, Long Short-Term Memory (LSTM) and Support Vector Regression (SVR), in predicting the stock prices of PT Bank Rakyat Indonesia (BBRI) using daily historical data from January 1, 2020, to January 10, 2025. The data were processed using a 60-day sliding window technique and normalized with MinMaxScaler. Model performance was evaluated using Mean Absolute Error (MAE), Mean Squared Error (MSE), and the coefficient of determination (R²) across five independent trials (5-fold trials). The evaluation results show that SVR outperforms in short-term prediction, with an average MAE of 0.0281, MSE of 0.0014, and R² of 0.9072. Meanwhile, LSTM records an average MAE of 0.0312, MSE of 0.0015, and R² of 0.8962, but achieves better performance in medium-term predictions, with a smaller average error of Rp228.02 compared to Rp242.52 from SVR. Both models demonstrate strong generalization capabilities on test data without signs of overfitting. Based on these findings, SVR is recommended for stable short-term forecasts, while LSTM is better suited for medium-term predictions involving complex trend patterns.This study aims to compare the performance of two machine learning algorithms, Long Short-Term Memory (LSTM) and Support Vector Regression (SVR), in predicting the stock prices of PT Bank Rakyat Indonesia (BBRI) using daily historical data from January 1, 2020, to January 10, 2025. The data were processed using a 60-day sliding window technique and normalized with MinMaxScaler. Model performance was evaluated using Mean Absolute Error (MAE), Mean Squared Error (MSE), and the coefficient of determination (R²) across five independent trials (5-fold trials). The evaluation results show that SVR outperforms in short-term prediction, with an average MAE of 0.0281, MSE of 0.0014, and R² of 0.9072. Meanwhile, LSTM records an average MAE of 0.0312, MSE of 0.0015, and R² of 0.8962, but achieves better performance in medium-term predictions, with a smaller average error of Rp228.02 compared to Rp242.52 from SVR. Both models demonstrate strong generalization capabilities on test data without signs of overfitting. Based on these findings, SVR is recommended for stable short-term forecasts, while LSTM is better suited for medium-term predictions involving complex trend patterns
Implementation of The Logistic Regression Algorithm to Analyze Poverty Factors in Aceh Province
Get PDFAceh Province continues to face a high poverty rate despite its abundant natural resources. This study aims to analyze the factors influencing poverty status in Aceh Province by applying a binary logistic regression algorithm. The research specifically focuses on an inferential analytical approach to reveal significant relationships among socioeconomic variables. Secondary data were obtained from the Aceh Provincial Statistics Agency (Badan Pusat Statistik/BPS) for the period 2019–2023. Inferential analysis was conducted using the entire dataset through the statsmodels library to identify variables that are statistically significant to poverty status. In addition, a classification approach was implemented using scikit-learn, with a data split between training data (2019–2022) and testing data (2023), yielding an accuracy of 0.70, precision of 0.81, recall of 0.70, F1-score of 0.66, and AUC of 0.69. These findings provide empirical evidence that improving access to education and equitable infrastructure development in densely populated areas can serve as effective policy focuses in efforts to alleviate poverty in Aceh Province
Clustering and Forecasting Implementation for Medical Consumables Stock Reccomendation
Get PDFManaging medical consumables (BMHP) in hospitals can be tricky because the demand often changes unpredictably. This study aims to help hospitals manage their BMHP stocks better by using two techniques: forecasting with Single Exponential Smoothing (SES) and grouping items using Agglomerative Hierarchical Clustering (AHC). SES is used to predict future needs based on previous usage, while AHC groups similar items based on how they\u27re used, which helps make the predictions more accurate. Before applying clustering, the prediction error was quite high, with a MAPE of 61.77% and an MAE of 18,769.80. After clustering, these numbers dropped to 10.06% and 3,987.45, showing a significant improvement. The clustering itself was strong, with a Silhouette Coefficient of 0.727, meaning the item groups made sense. Each group of items got different stock suggestions. Items with high and unstable demand were advised to keep extra safety stock. Items with uncertain patterns needed a more flexible buffer stock. For items with stable use, average trends over the last few months were enough to guide stock planning. This approach helps hospitals avoid both overstock and stockouts by giving more accurate and tailored recommendations. Although this study only used data from one hospital, the results show that combining SES and AHC can make stock management smarter and more efficient
Browser-Based Detection of Harmful Content with Deep Learning Model
Get PDFThis study presents a browser extension that detects harmful content on both web pages and TikTok using a deep learning-based approach. The core model employs a Bidirectional Long Short-Term Memory (BiLSTM) network for multi-label classification, targeting six categories: Toxic, Severe Toxic, Obscene, Threat, Insult, and Identity Hate. The dataset combines 13,057 labeled samples from a public Kaggle dataset (2021) and 2,884 manually labeled tweets scraped from Twitter (X) between October–November 2024. Three feature extraction methods were tested: learned embeddings, FastText, and Word2Vec. The BiLSTM model architecture includes one embedding layer, a 32-unit bidirectional LSTM, three dense layers (128,256,128) using ReLU activation, and a six-unit sigmoid output layer. The model was trained using the Adam optimizer and binary cross-entropy loss, with early stopping applied after five stagnant validation checks across a maximum of 200 epochs. While the FastText-based model showed the best performance, the final deployed model used learned embeddings in Scenario 1 due to its smaller size (1.6M parameters) and near-optimal performance (Recall: 0.9786; Hamming Loss: 0.0052). The extension also integrates Whisper ASR for detecting harmful speech in video-based platforms like TikTok and supports five customizable censorship filters. User evaluation via Customer Satisfaction Score (CSAT) indicated strong acceptance, with 95.45% rating the user experience as Excellent, 84.09% confirming detection relevance, and 79.55% rating the system performance as Good. This highlights the extension’s effectiveness in promoting safer digital interaction across text and audiovisual content
Optimasi Kualitas Animasi “Generate” melalui Video Referensi Menggunakan Metode Pose to Pose
Get PDFThis study aims to optimize the quality of the 3D animated short film “Generate” by applying the pose-to-pose animation method supported by the effective use of reference videos. Previously, the animation was done using the straight-ahead technique, which was considered ineffective in highlighting key poses, timing, and arcs. Through an experimental qualitative approach, the animation was improved using the pose-to-pose method, which included analyzing reference videos to determine the correct key poses, breakdowns, and in-betweens. The resulting animation was then validated by professional animation practitioners based on aspects of timing, arcs, and overall visual quality. The findings showed a significant improvement in the clarity of the animation, arcs, and character poses. The pose-to-pose method is also considered effective for novice animators because it provides a more structured workflow to support staging and storytelling. However, challenges remain in the ability to accurately understand and apply references, requiring more in-depth training. This study offers a practical framework that can bridge the animation learning process in an academic environment with professional industry standards.Penelitian ini bertujuan untuk mengoptimalkan kualitas film pendek animasi 3D "Generate" dengan menerapkan metode animasi pose-to-pose yang didukung oleh penggunaan video referensi yang efektif. Sebelumnya, animasi dilakukan dengan teknik straight-ahead, yang dianggap kurang efektif dalam menonjolkan pose-pose utama, pengaturan waktu, dan lengkungan. Melalui pendekatan kualitatif eksperimental, animasi ditingkatkan menggunakan metode pose-to-pose, yang mencakup analisis video referensi untuk menentukan pose-pose utama yang tepat, breakdown, dan in-between. Animasi yang dihasilkan kemudian divalidasi oleh praktisi animasi profesional berdasarkan aspek pengaturan waktu, lengkungan, dan kualitas visual secara keseluruhan. Temuan menunjukkan peningkatan yang signifikan dalam kejelasan animasi, lengkungan, dan pose karakter. Metode pose-to-pose juga dianggap efektif bagi animator pemula karena menyediakan alur kerja yang lebih terstruktur untuk mendukung staging dan penceritaan. Namun, tantangan yang ada adalah kemampuan untuk memahami dan menerapkan referensi secara akurat, sehingga membutuhkan pelatihan yang lebih mendalam. Penelitian ini menawarkan kerangka kerja praktis yang dapat menjembatani proses pembelajaran animasi di lingkungan akademis dengan standar industri profesional
Prototype of Implementation of Smart Contract for Blockchain-Based Document Storage
Get PDFData, including digital and physical documents, is a valuable asset often vulnerable to forgery, theft, and reliance on centralized servers, which are costly and prone to failure. This study develops a prototype of a decentralized document storage application by combining blockchain and the InterPlanetary File System (IPFS). The system is designed as a web-based decentralized application (DApp), integrating Ethereum smart contracts to immutably record document metadata and access history, while the actual files are stored in IPFS and identified using unique Content Identifiers (CIDs). User interactions are facilitated through MetaMask for authentication and transaction approval. The system is developed using the Waterfall methodology. Functional testing is conducted through unit tests using Ganache as a local Ethereum blockchain, and the smart contract is also deployed to the Sepolia Ethereum testnet. The results show that the system successfully stores documents via IPFS and records metadata and access activities transparently on the blockchain. Access and download tracking features enhance document accountability. This solution provides a secure, efficient, and transparent alternative to centralized document storage and contributes to the advancement of distributed digital archiving systems
Determining Eligibility for Smart Indonesia Program (PIP) Recipients Using the Backpropagation Method
Get PDFThe government provides financial assistance, educational opportunities, and expands access for students from poor or vulnerable families through the Smart Indonesia Program (PIP). At Madrasah Ibtidaiyah Negeri 20 Bireuen, the selection process for underprivileged students is still carried out manually by homeroom teachers by collecting data on students and their parents. This study aims to design, implement, and evaluate a classification method using the Backpropagation Neural Network to determine the eligibility of PIP scholarship recipients. The dataset consists of 309 entries, comprising 217 training data and 92 testing data, collected from MIN 20 Bireuen students between 2021 and 2023. The attributes used include father\u27s occupation, mother\u27s occupation, father\u27s income, mother\u27s income, number of dependents, number of vehicles, home ownership status, and card ownership status. Prior to training, the data were normalized using Min-Max scaling. The model was built with one hidden layer using a hard-limit activation function and a learning rate of 0.01. The classification results are categorized as "Eligible" and "Not Eligible". The model achieved an accuracy of 98%, precision of 100%, recall of 95%, and F1-score of 97%