Jurnal Politeknik Negeri Batam (PoliBatam)
Not a member yet
3001 research outputs found
Sort by
Utilization of Machine Learning for Predicting Corrosion Inhibition by Quinoxaline Compounds
Get PDFCorrosion is a significant issue in both industrial and academic sectors, with widespread negative impacts on various aspects, including economics and safety. To address this problem, the use of corrosion inhibitors has proven effective. This study explores the application of Machine Learning (ML) methods based on Quantitative Structure-Properties Relationship (QSPR) to develop a predictive model for the efficiency of quinoxaline compounds as corrosion inhibitors. By conducting a comparative analysis among three algorithms: AdaBoost Regressor (ADB), Gradient Boosting Regressor (GBR), and Extreme Gradient Boosting Regressor (XGBR), and optimizing parameters through hyperparameter tuning using Grid Search and Random Search, this research demonstrates that the XGBR model yields the most superior prediction results. The XGBR optimized with hyperparameter tuning using Grid Search achieved the highest R² value of 0.970 and showed the lowest RMSE, MSE, MAD, and MAPE values of 0.368, 0.135, 0.119, and 0.273, respectively, indicating high predictive accuracy. These results are expected to contribute to the development of more effective methods for identifying corrosion inhibitor candidates.Corrosion is a significant issue in both industrial and academic sectors, with widespread negative impacts on various aspects, including economics and safety. To address this problem, the use of corrosion inhibitors has proven effective. This study explores the application of Machine Learning (ML) methods based on Quantitative Structure-Properties Relationship (QSPR) to develop a predictive model for the efficiency of quinoxaline compounds as corrosion inhibitors. By conducting a comparative analysis among three algorithms: AdaBoost Regressor (ADB), Gradient Boosting Regressor (GBR), and Extreme Gradient Boosting Regressor (XGBR), and optimizing parameters through hyperparameter tuning using Grid Search and Random Search, this research demonstrates that the XGBR model yields the most superior prediction results. The XGBR optimized with hyperparameter tuning using Grid Search achieved the highest R² value of 0.970 and showed the lowest RMSE, MSE, MAD, and MAPE values of 0.368, 0.135, 0.119, and 0.273, respectively, indicating high predictive accuracy. These results are expected to contribute to the development of more effective methods for identifying corrosion inhibitor candidates
Sentiment Analysis on Google Reviews Using Naïve Bayes, K-Nearest Neighbors, and Logistic Regression to Improve Novotel Services
Get PDFThe application of artificial intelligence (AI) has been widely used in various industrial sectors, including the hospitality industry. One of the applications that is widely used in the hospitality industry is sentiment analysis. Sentiment analysis is carried out by analyzing feedback data from hotel guests or customers. The results of this sentiment analysis are important for decision makers to improve and improve their services. This study aims to obtain sentiment analysis results from Novotel hotel Google reviews based on machine learning by comparing three algorithms, namely Naïve Bayes, K-Nearest Neighbors (KNN), and Logistic Regression. The stages carried out in this study are data collection, data labeling, exploratory data analysis (EDA), data preprocessing, text representation, data sharing, modeling, model training, model evaluation, selection of the most accurate model, visualization of the most accurate model, interpretation of results and writing research reports. The dataset used was 1200 reviews, only 1190 reviews were used in the analysis. From the training results, the model produced by the Logistic Regression algorithm was the most accurate, namely 94.54% with unigrams (n = 1). Here are the results of each category, positive as many as 723 reviews (60.76%), negative as many as 218 reviews (18.32%), and neutral as many as 249 reviews (20.92%). Thus, most of the sentiment towards the service is positive, but some services need to be fixed and improved for customer satisfaction. The next research, the research area is expanded and the use of Deep Learning.The application of artificial intelligence (AI) has been widely used in various industrial sectors, including the hospitality industry. One of the applications that is widely used in the hospitality industry is sentiment analysis. Sentiment analysis is carried out by analyzing feedback data from hotel guests or customers. The results of this sentiment analysis are important for decision makers to improve and improve their services. This study aims to obtain sentiment analysis results from Novotel hotel Google reviews based on machine learning by comparing three algorithms, namely Naïve Bayes, K-Nearest Neighbors (KNN), and Logistic Regression. The stages carried out in this study are data collection, data labeling, exploratory data analysis (EDA), data preprocessing, text representation, data sharing, modeling, model training, model evaluation, selection of the most accurate model, visualization of the most accurate model, interpretation of results and writing research reports. The dataset used was 1200 reviews, only 1190 reviews were used in the analysis. From the training results, the model produced by the Logistic Regression algorithm was the most accurate, namely 94.54% with unigrams (n = 1). Here are the results of each category, positive as many as 723 reviews (60.76%), negative as many as 218 reviews (18.32%), and neutral as many as 249 reviews (20.92%). Thus, most of the sentiment towards the service is positive, but some services need to be fixed and improved for customer satisfaction. The next research, the research area is expanded and the use of Deep Learning
Pengklasifikasian Warna dan Bentuk Produk Menggunakan Kamera ELP- USB8MP02G-MFV dengan Berbasis YOLOV7
No full textThe development of artificial intelligence technology allows the system to detect various objects. In the research on the classification of color and shape of products using the ELP-USB8MP02G-MFV camera based on YOLOV7, it aims to modify the conveyor on the molding machine. Because the conveyor only has the function of distributing goods from the molding machine to the bin and the length of time used to wait for the bin to be full is the reason why this conveyor is modified. Modifications are made by adding a camera that has been connected to the Raspberry Pi 4B on the conveyor, the camera functions to take pictures of passing product objects then the image is detected by the system on the Raspberry Pi 4B so that this conveyor machine can classify the objects produced by the molding machine. The system detects objects using the YOLOv7 algorithm. This study was carried out with three tests, namely object model detection testing, color detection testing and program and relay output testing where 98.11% was for object model detection testing, 97.37% for color detection and 100% for program and relay output testing. The results of this research will contribute to the development of object detection, especially product object detection and the results of molding machines.Semakin berkembangnya teknologi kecerdasan buatan memungkinkan sistem dapat mendeteksi berbagai objek.Pada penelitian pengklasifikasian warna dan bentuk produk menggunakan kamera ELP-USB8MP02G-MFV dengan berbasis YOLOV7 bertujuan untuk untuk memodifikasi konveyor yang berada pada mesin molding.Karena konveyor hanya memiliki fungsi untuk menyalurkan barang dari mesin molding menuju bin dan lamanya waktu yang digunakan untuk menunggu bin penuh menjadi alasan mengapa konveyor ini dimodifikasi. Modifikasi dilakukan dengan cara menambahkan kamera yang telah dihubungkan dengan Raspberry Pi 4B pada konveyor, kamera tersebut berfungsi untuk mengambil gambar objek produk yang lewat kemudian gambar tersebut dideteksi oleh sistem pada Raspberry Pi 4B sehingga mesin konveyor ini dapat mengklasifikasikan objek hasil produk mesin molding. Sistem ini mendeteksi objek menggunakan algoritma YOLOv7. Penelitian ini dilakukan dengan tiga pengujian yaitu pengujian pendeteksian model objek, pengujian pendeteksi warna dan pengujian output program dan relay dimana 98.11% untuk pengujian pendeteksian model objek, 97.37% untuk pendeteksian warna dan 100 % untuk pengujian output program dan relay. Hasil penelitian ini akan memberikan kontribusi terkait pegembangan pendeteksian objek, khususnya pendeteksian objek produk dan hasil dari mesin molding
Rancang Bangun Alat Penetas Telur dengan Sistem Pengendali Suhu dan Kelembapan Menggunakan Metode Fuzzy dan Monitoring Berbasis IoT: Rancang Bangun Alat Penetas Telur dengan Sistem Pengendali Suhu dan Kelembapan Menggunakan Metode Fuzzy dan Monitoring Berbasis IoT
No full textThe egg hatching process is an important stage in the poultry farming industry that requires optimal temperature and humidity control. This research develops an automatic egg hatcher equipped with Mamdani fuzzy logic-based control and Internet of Things (IoT)-based monitoring. The fuzzy system is used to precisely regulate temperature and humidity, while IoT technology enables real-time monitoring through a digital platform. The test results show high accuracy with an average error of 0.78% for temperature and 1.08% for humidity. IoT monitoring has reliable performance with a data transmission delay of 5-7 seconds. In addition, it achieved 71,4% hatchability, proving its effectiveness in creating optimal environmental conditions for hatching. It provides a modern solution that improves the efficiency, productivity, and quality of hatching eggs, and makes it easier for farmers to monitor and control the process automatically.Proses penetasan telur merupakan tahap penting dalam industri peternakan unggas yang memerlukan pengendalian suhu dan kelembapan secara optimal. Penelitian ini mengembangkan alat penetas telur otomatis yang dilengkapi dengan pengendalian berbasis logika fuzzy Mamdani dan monitoring berbasis Internet of Things (IoT). Sistem fuzzy digunakan untuk mengatur suhu dan kelembapan secara presisi, sementara teknologi IoT memungkinkan pemantauan real-time melalui website. Hasil pengujian menunjukkan akurasi tinggi dengan rata-rata error sebesar 0,78% untuk suhu dan 1,08% untuk kelembapan. Monitoring IoT memiliki performa andal dengan delay pengiriman data 5-7 detik. Selain itu, alat ini berhasil mencapai daya tetas 71,4%, membuktikan efektivitasnya dalam menciptakan kondisi lingkungan optimal untuk penetasan. Alat ini memberikan solusi modern yang meningkatkan efisiensi, produktivitas, dan kualitas hasil penetasan telur, serta mempermudah peternak dalam memantau dan mengontrol proses secara otomatis
Comparative Analysis of VGG16 and ResNet50 Model Performence in Cardiac ECG Image Classification
No full textThis study systematically evaluates and compares the effectiveness of two deep learning architectures, VGG16 and ResNet50, in automating electrocardiogram (ECG) image classification for cardiac condition diagnosis. The dataset was obtained from a public source and consists of 2,898 color ECG images converted from raw signals, categorized into four classes: Abnormal Heartbeat, Myocardial Infarction, Normal Individual, and History of Heart Attack. The data underwent preprocessing steps including resizing to 224×224 pixels, pixel normalization to a 0–1 range, label encoding, one-hot encoding, and an 80:20 split for training and testing. Transfer learning was applied using feature representations from the VGG16 and ResNet50 models, employing the Adam optimizer and categorical cross-entropy loss function. To enhance training efficiency and prevent overfitting, early stopping was implemented based on validation loss performance. Model performance was evaluated using accuracy, precision, recall, and F1-score metrics. The results showed that VGG16 achieved 95% accuracy with a loss of 0.1522, precision of 95%, recall of 94%, and F1-score of 94%. In contrast, ResNet50 attained 81% accuracy with a loss of 0.5730, precision of 82%, recall of 79%, and F1-score of 80%. These findings indicate that, within the context of this study, VGG16 consistently outperformed ResNet50 across all evaluation metrics in the ECG image classification task. Therefore, the application of transfer learning using the VGG16 model demonstrates strong potential as an effective approach for AI-based ECG image classification systems.This study systematically evaluates and compares the effectiveness of two deep learning architectures, VGG16 and ResNet50, in automating electrocardiogram (ECG) image classification for cardiac condition diagnosis. The dataset was obtained from a public source and consists of 2,898 color ECG images converted from raw signals, categorized into four classes: Abnormal Heartbeat, Myocardial Infarction, Normal Individual, and History of Heart Attack. The data underwent preprocessing steps including resizing to 224×224 pixels, pixel normalization to a 0–1 range, label encoding, one-hot encoding, and an 80:20 split for training and testing. Transfer learning was applied using feature representations from the VGG16 and ResNet50 models, employing the Adam optimizer and categorical cross-entropy loss function. To enhance training efficiency and prevent overfitting, early stopping was implemented based on validation loss performance. Model performance was evaluated using accuracy, precision, recall, and F1-score metrics. The results showed that VGG16 achieved 95% accuracy with a loss of 0.1522, precision of 95%, recall of 94%, and F1-score of 94%. In contrast, ResNet50 attained 81% accuracy with a loss of 0.5730, precision of 82%, recall of 79%, and F1-score of 80%. These findings indicate that, within the context of this study, VGG16 consistently outperformed ResNet50 across all evaluation metrics in the ECG image classification task. Therefore, the application of transfer learning using the VGG16 model demonstrates strong potential as an effective approach for AI-based ECG image classification systems
Modeling Productive Land Determination Using Entropy-Mabac Method Based on Multicriteria Data in Central Java Province
No full textCentral Java Province has a diversity of land use characteristics that reflect the potential as well as challenges in regional development, so that optimization of productive land is important to support economic growth, community welfare, and environmental sustainability. For this reason, this research was conducted with an objective approach using the Entropy method in determining the weight of each criterion based on actual data variations, as well as the Multi-Attributive Border Approximation Area Comparison (MABAC) method to systematically evaluate and rank the level of land productivity in 35 districts/cities. The results of the analysis show that Demak, Brebes, and Rembang districts ranked the highest in land productivity with the highest score of 0.249, while Wonogiri and Banjarnegara districts ranked the lowest with scores of -0.392 and -0.234. Validation using the Spearman Rank test resulted in a correlation coefficient of 0.82, indicating strong agreement between the method results and historical data. The findings show that the combination of Entropy and MABAC methods is effective in determining productive land, and the results are relevant as a basis for formulating sustainable land use policies, including recommendations for irrigation development, farmland protection, and strengthening spatial policies for low productivity areas
A Design and Implementation of a 3-Axis UAV Drone Gimbal Rig for Testing Stability and Performance Parameters in the Laboratory
No full textThis study designs a 3-axis UAV gimbal rig for testing stability and performance before deployment in real-world flight conditions. The gimbal rig simulates the vertical, lateral, and longitudinal axes to ensure reliable operation in various scenarios. Made from lightweight aluminum alloy, the structure minimizes vibrations and maintains rigidity during testing. For precise motion tracking, each axis is equipped with an LPD3806- 600BM-G5 rotary encoder, offering accurate feedback on movement. The Arduino Nano processes the encoder data, displaying real-time results on a 16x2 LCD with an I2C interface for easy monitoring. Additionally, a push-button system enables users to switch between different readings for each axis. This setup aids researchers in analyzing UAV dynamics and refining both firmware and hardware. Future enhancements may include wireless data logging and integration of machine learning techniques to predict maintenance needs, further supporting UAV stability testing in various applications, including aerospace, defense, and commercial use
Application of Convolutional Neural Network (CNN) Algorithm with ResNet-101 Architecture for Monkey Pox Detection in Human
No full textMonkeypox is a zoonotic disease that has spread to various countries, including Indonesia. It is transmitted through direct contact with skin lesions, respiratory droplets, or contaminated objects. Early and accurate detection is crucial to reduce the risk of transmission and improve treatment effectiveness. This study aims to detect monkeypox using a Convolutional Neural Network (CNN) with the ResNet-101 architecture. The pre-processing steps include normalization and resizing of images to 224×224 pixels. The model is trained using the Adam optimizer, categorical crossentropy loss function, and an adaptive learning rate reduction. Evaluation results show that the model achieved an accuracy of 94%, with a precision of 0.92, recall of 0.92, and an F1-score of 0.92. The model is capable of classifying images effectively, although some misclassifications still occur. This system is intended to function as an initial image-based screening tool, but its results should be confirmed through clinical diagnosis and laboratory testing to ensure accuracy.Monkeypox is a zoonotic disease that has spread to various countries, including Indonesia. It is transmitted through direct contact with skin lesions, respiratory droplets, or contaminated objects. Early and accurate detection is crucial to reduce the risk of transmission and improve treatment effectiveness. This study aims to detect monkeypox using a Convolutional Neural Network (CNN) with the ResNet-101 architecture. The pre-processing steps include normalization and resizing of images to 224×224 pixels. The model is trained using the Adam optimizer, categorical crossentropy loss function, and an adaptive learning rate reduction. Evaluation results show that the model achieved an accuracy of 94%, with a precision of 0.92, recall of 0.92, and an F1-score of 0.92. The model is capable of classifying images effectively, although some misclassifications still occur. This system is intended to function as an initial image-based screening tool, but its results should be confirmed through clinical diagnosis and laboratory testing to ensure accuracy
Scenario-Based Association Rule Mining in Veterinary Services Using FP-Growth: Differentiating Clinical and Customer-Driven Patterns
No full textVeterinary clinics routinely generate transactional data that contain valuable information about both operational workflows and customer preferences. This study aims to differentiate between procedural and customer-driven service patterns by applying the FP-Growth association rule mining algorithm to 1,000 anonymized transactions comprising 94 unique items, collected from a veterinary clinic in West Java, Indonesia, during 2023. Two distinct analytical scenarios were constructed: Scenario 1 includes all services (procedural and customer-driven), while Scenario 2 excludes procedural items such as “Vet” and “Visit Dokter” to focus solely on client-initiated behaviors. Data preprocessing involved aggregating transaction items into a market basket format suitable for frequent pattern mining. The FP-Growth algorithm was employed to extract association rules, evaluated using support, confidence, and lift metrics. Results from Scenario 1 revealed rule patterns reflective of standard clinical protocols and operational dependencies, informing bundled service packages and inventory management. In contrast, Scenario 2 uncovered customer-driven associations, highlighting opportunities for personalized promotions and service innovation. The comparative analysis demonstrates the utility of scenario-based association rule mining for both operational optimization and customer engagement. While the findings provide actionable insights for clinic management, further validation with practitioners and implementation in multi-clinic settings are recommended to confirm real-world applicability and enhance generalizability
Strategi Perancangan Dan Penerapan Konten Instagram @Helloskinhouse Untuk Meningkatkan Engagement
Get PDFThe growth of the beauty industry in Batam demands business owners, such as Hello Skin, implement effective digital marketing strategies, particularly through the Instagram social media platform. This study aims to design and implement a content marketing strategy for the Instagram account @helloskinhouse to enhance customer engagement. The development method used is design thinking, comprising five stages, empathize, define, ideate, prototype, and test. Data were collected through observations, interviews, and the distribution of questionnaires to 202 respondents, who are followers of @helloskinhouse. The results of the study indicate that the implementation of the content marketing strategy, which includes a variety of content types, persuasive copywriting, appealing visuals, the use of relevant hashtags, and optimal posting times, significantly increased audience interactions by 2,500%. The content marketing index reached 86.20%, and customer engagement achieved 87.22%, both of which are categorized as excellent. Simple linear regression and partial tests (t-tests) demonstrated that content marketing has a positive and significant effect on customer engagement (significance value < 0.05). Thus, a well-structured content marketing strategy has been proven effective in increasing audience engagement and building brand trust on the @helloskinhouse Instagram account amid the competitive beauty industry.Pertumbuhan industri kecantikan di Batam menuntut para pelaku usaha seperti Hello Skin untuk menerapkan strategi pemasaran digital yang efektif, khususnya melalui platform media sosial Instagram. Penelitian ini bertujuan untuk merancang dan mengimplementasikan strategi pemasaran konten pada akun Instagram @helloskinhouse guna meningkatkan customer engagement. Metode pengembangan yang digunakan adalah design thinking yang terdiri dari lima tahap, yaitu empathize, define, ideate, prototype, dan test. Pengumpulan data dilakukan melalui observasi, wawancara, dan penyebaran kuesioner kepada 202 responden yang merupakan followers akun @helloskinhouse. Hasil penelitian menunjukkan bahwa penerapan strategi pemasaran konten yang meliputi variasi jenis konten, copywriting yang persuasif, visual yang menarik, penggunaan hashtag yang relevan, dan waktu posting yang optimal mampu meningkatkan interaksi audiens secara signifikan hingga 2.500%. Indeks pemasaran konten mencapai 86,20% dan customer engagement mencapai 87,22%, keduanya masuk dalam kategori sangat baik. Regresi linier sederhana dan uji parsial (uji-t) menunjukkan bahwa pemasaran konten memiliki efek positif dan signifikan terhadap keterlibatan pelanggan (nilai signifikansi < 0,05). Dengan demikian, strategi pemasaran konten yang terstruktur dengan baik terbukti efektif dalam meningkatkan keterlibatan audiens dan membangun kepercayaan merek pada akun Instagram @helloskinhouse di tengah persaingan industri kecantikan