Jurnal Politeknik Negeri Batam (PoliBatam)
Not a member yet
3001 research outputs found
Sort by
Optimization of Decision Tree Algorithm for Chronic Kidney Disease Classification Based on Particle Swarm Optimization (PSO)
Get PDFThe body\u27s most important vital organ is the kidney. The kidneys are responsible for maintaining acid and alkaline balance, regulating blood pressure, and filtering blood to prevent the accumulation of metabolic waste in the body. However, chronic kidney disease does not always show symptoms and signs but can progress to kidney failure. Algorithm-based predictive methods in data processing show great potential in the health field to predict various diseases, one of which is kidney disease. One of the techniques in data mining is classification. One of the classification algorithms in data mining that is often used to detect diseases is Decision Tree. In this study, it is expected that by combining these two methods, it will make a new contribution to the Decision Tree algorithm that is optimized with Particle Swarm Optimization (PSO) for the selection of relevant features, and improve the weaknesses in the model to improve more accurate predictions. By performing feature selection with the Particle Swarm Optimization (PSO) algorithm, it is shown that the use of Particle Swarm Optimization (PSO) can improve the accuracy and performance of the Decision Tree algorithm in the chronic kidney disease classification process. The accuracy of the Decision Tree algorithm with feature selection using Particle Swarm Optimization (PSO) is higher, reaching 0.967%, compared to the accuracy of Decision Tree without Particle Swarm Optimization (PSO) feature selection which is only 0.95%. This shows that Particle Swarm Optimization (PSO) is effective in selecting relevant features so that it can significantly improve model performance.The body\u27s most important vital organ is the kidney. The kidneys are responsible for maintaining acid and alkaline balance, regulating blood pressure, and filtering blood to prevent the accumulation of metabolic waste in the body. However, chronic kidney disease does not always show symptoms and signs but can progress to kidney failure. Algorithm-based predictive methods in data processing show great potential in the health field to predict various diseases, one of which is kidney disease. One of the techniques in data mining is classification. One of the classification algorithms in data mining that is often used to detect diseases is Decision Tree. In this study, it is expected that by combining these two methods, it will make a new contribution to the Decision Tree algorithm that is optimized with Particle Swarm Optimization (PSO) for the selection of relevant features, and improve the weaknesses in the model to improve more accurate predictions. By performing feature selection with the Particle Swarm Optimization (PSO) algorithm, it is shown that the use of Particle Swarm Optimization (PSO) can improve the accuracy and performance of the Decision Tree algorithm in the chronic kidney disease classification process. The accuracy of the Decision Tree algorithm with feature selection using Particle Swarm Optimization (PSO) is higher, reaching 0.967%, compared to the accuracy of Decision Tree without Particle Swarm Optimization (PSO) feature selection which is only 0.95%. This shows that Particle Swarm Optimization (PSO) is effective in selecting relevant features so that it can significantly improve model performance
Web GIS Based Benthic Habitat Mapping Update Supports Smart Island Lemukutan
Get PDFBenthic habitats are important for the quality of life and global climate. Systematic and efficient information is important for the monitoring, mapping, and recording of aquatic bottom habitats, thus providing a habitat database. In the last decade, object-based image analysis (OBIA) has been accepted as an effective method for extracting and classifying information from high spatial resolution satellite imagery. Our study\u27s goal is to use WebGIS to combine coral reef monitoring data from Lemukutan Island and find out how much coral cover there is on the island using the smart island WebGIS. This study took place from August 6th to August 13th, 2024, and used a total of 1097 field points to show where all the benthic habitats and Sentinel 2A image data sources were located. The research results obtained the extent of shallow water benthic habitat classification with different variations in each habitat class. The Rock Class covers an area of "‹"‹41,940 ha, the mixed class 2,409 ha, the coral class 130,340 ha, the dead coral class 49,249 ha, the macroalgae class 2,840 ha, and the sand class 12,140 ha. The overall accuracy (OA) results for the waters of Lemukutan Island obtained the highest value, namely 89.5833%, using the SVM algorithm. Regular monitoring of coral reefs can help update Lemukutan Island Smart Island data to become a catalyst in realizing a smart island ecosystem in West Kalimantan Province by presenting benthic habitat maps through web GIS services and realizing technology development for coastal areas and small islands
Desain dan Implementasi Kotak Pintar (Kopin.COD) Penerima Paket COD dan Paket Non COD Berbasis IoT
No full textThis study employs the black-box method and qualitative user analysis. Data collection is conducted through surveys targeting users of the developed tool, aiming to identify common issues in online shopping with the Cash on Delivery (COD) payment method. These issues include customers not being at home when the courier delivers the package, resulting in repeated delivery attempts and delays in receiving packages. Based on this identification, the team determined a solution: creating a smart COD package receiver box integrated with a mobile application, equipped with barcode verification and a surveillance camera. The results show that 98% of users are satisfied with the convenience provided by the Kopin.COD system and feel significantly assisted by this technology. Additionally, survey results from customers or package recipients indicate that 100% of respondents are satisfied with the ease and convenience of use. Penelitian ini menggunakan metode analisa kualitatif pengguna. Pengumpulan data menggunakan survei terhadappengguna terkait alat yang sudah dibuat dengan mengidentifikasi permasalahan yang sering terjadi dalam belanjaonline dengan metode pembayaran Cash On Delivery (COD), seperti pelanggan yang tidak berada di rumah saatkurir mengantar paket, yang mengakibatkan pengulangan pengantaran paket dan keterlambatan dalam menerimapaket. Berdasarkan identifikasi tersebut, tim menentukan solusi yang ingin dicapai yaitu menciptakan produkkotak pintar penerima paket COD yang terintegrasi dengan aplikasi mobile, dilengkapi dengan verifikasi barcodedan kamera pengawas. Hasil yang didapat 98% di antaranya merasa puas dengan kemudahan yang diberikan olehsistem Kopin.COD dan merasa sangat terbantu dengan adanya teknologi ini. Selain itu, hasil survei yangdilakukan terhadap pelanggan atau penerima paket menunjukkan bahwa 100% responden merasa puas dengankemudahan serta kenyamanan penggunaan
Akuisisi Data dan Kalibrasi pada Sistem Can Satellite (Cansat)
No full textThis study aims to develop and test a data acquisition and calibration system for CanSat, a small nanosatellite designed as a cost-effective solution for collecting geographical and spatial data. CanSat collects data using BME-280, MPU-9250, and NEO-6M sensors, then transmits it via XBee to a GUI for controlling the release of the Heat Shield, parachute, and legs. The data is analyzed using statistical methods and transmitted via XBee S2C. The results show that the BME-280 has a standard error of 0.35 and a deviation of 0.77 for temperature, and a standard error of 0.06 with a deviation of 0.13 for air pressure. The altitude measurement has a standard error of 0.02 and a deviation of 0.05. Meanwhile, MPU-9250 has an average standard error of 0.34 and a deviation of 0.77 for angle orientation measurement. The design and testing of the leg control system conclude that it functions at 80% efficiency. Penelitian ini bertujuan untuk mengembangkan dan menguji sistem akuisisi data dan kalibrasi pada Cansat, sebuahnanosatelit berukuran kecil yang dirancang sebagai solusi hemat biaya untuk pengumpulan data geografis danspasial. CanSat mengukur data menggunakan sensor BME-280, MPU-9250, dan NEO-6M, lalu mengirimkannyamelalui XBee ke GUI untuk mengontrol pelepasan Heat Shield, parasut, dan kaki. Data diuji dengan metodestatistik dan dikirim menggunakan XBee S2C. Hasilnya, BME-280 memiliki standar error 0,35 dan deviasi 0,77untuk suhu, serta standar error 0,06 dan deviasi 0,13 untuk tekanan udara. Pengukuran ketinggian memiliki standarerror 0,02 dan deviasi 0,05. Sementara itu, MPU-9250 dalam pengukuran orientasi sudut memiliki rata-ratastandar error 0,34 dan deviasi 0,77. hasil perancangan dan pengujian pada sistem kontrol bedirinya kaki dapatdisimpulkan bahwa sistem kontrol berfungsi 80%
Analisis Efektivitas Papan Reklame Berbasis IoT Dengan Metode Faster R-CNN
No full textPapan reklame merupakan media untuk mempublikasikan produk atau jasa, namun efektivitasnya sulit diukur karena tidak diketahui jumlah orang yang melihatnya. Oleh karena itu, penelitian ini mengembangkan sistem berbasis metode Faster R-CNN untuk mendeteksi jumlah viewers papan reklame. Sistem ini bertujuan membantu pemakai jasa iklan dalam menilai efektivitas pemasangan iklan serta memberi nilai tambah bagi penyedia jasa dengan menyediakan data jumlah viewers sebagai daya tarik layanan. Hasil pengujian menunjukkan sistem mampu mendeteksi wajah dengan akurasi 56,74%, motor 76,47%, dan mobil 93,94%. Beberapa faktor yang mempengaruhi akurasi deteksi antara lain pencahayaan, jarak, resolusi kamera, serta kesesuaian dataset dengan lingkungan implementasi. Dengan adanya sistem ini, pemakai jasa dapat menentukan lokasi pemasangan iklan yang lebih strategis berdasarkan data real-time, sementara penyedia jasa dapat meningkatkan daya tarik layanan pemasangan iklan dengan data jumlah viewers sebagai nilai jual. Teknologi ini diharapkan mampu meningkatkan efektivitas pemasaran melalui papan reklame secara lebih akurat, efisien, dan terukur.Papan reklame merupakan media untuk mempublikasikan produk atau jasa, namun efektivitasnya sulit diukur karena tidak diketahui jumlah orang yang melihatnya. Oleh karena itu, penelitian ini mengembangkan sistem berbasis metode Faster R-CNN untuk mendeteksi jumlah viewers papan reklame. Sistem ini bertujuan membantu pemakai jasa iklan dalam menilai efektivitas pemasangan iklan serta memberi nilai tambah bagi penyedia jasa dengan menyediakan data jumlah viewers sebagai daya tarik layanan. Hasil pengujian menunjukkan sistem mampu mendeteksi wajah dengan akurasi 56.74%, motor 76.47%, dan mobil 93.94%. Beberapa faktor yang mempengaruhi akurasi deteksi antara lain pencahayaan, jarak, resolusi kamera, serta kesesuaian dataset dengan lingkungan implementasi. Dengan adanya sistem ini, pemakai jasa dapat menentukan lokasi pemasangan iklan yang lebih strategis berdasarkan data real-time, sementara penyedia jasa dapat meningkatkan daya tarik layanan pemasangan iklan dengan data jumlah viewers sebagai nilai jual. Teknologi ini diharapkan mampu meningkatkan efektivitas pemasaran melalui papan reklame secara lebih akurat, efisien, dan terukur
IoT-Based Smoking Violation Detection System Equipped with Object Detection Using YOLOv5s Algorithm
No full textSmoking is a common habit in Indonesia. The Indonesian government has implemented regulations on smoke-free areas, but violations of the smoke-free policy still often occur. Previous studies have developed smoking violation detection system based on the MQ sensor. However, the smoking violation detection system based only on the MQ sensor is less reliable because the detected gas could come from other sources. Therefore, this study discusses a smoking violation detection system that can automatically verify smoking violation activities using the MQ-7 sensor, MQ-135 sensor, and the YOLOv5s algorithm. The MQ-7 sensor that has been calibrated to detect CO in ppm units achieved an accuracy level of 89.84%. The MQ-135 sensor also has successfully detected ammonia and toluene in cigarette smoke in ppm units. The trained YOLOv5s algorithm achieved an average Precision of 91.9%, Recall 83.7%, F1-Score 87.6%, and mAP50 88.3%. The system is equipped with a speaker that will sound automatically after a verified smoking violation occurs and Telegram notifications in the form of text messages and images
Utilization of EfficientNet-B0 to Identify Oncomelania Hupensis Lindoensis as a Schistosomiasis Host
No full textSchistosomiasis caused by the Schistosoma japonicum worm is a significant health problem in Indonesia, especially in endemic areas such as the Napu Plateau and Bada Plateau. The main problem in controlling this disease is the difficulty in rapid and accurate identification of Oncomelania hupensis lindoensis snails as intermediate hosts of the parasite. This research aims to develop an artificial intelligence-based system that can efficiently identify the snail species. The stages of this research include collecting snail image data from the Central Sulawesi Provincial Health Office, consisting of 2100 images covering seven snail species, then processed through preprocessing and augmentation stages. The model applied was EfficientNet-B0. The results showed that the EfficientNet-B0 model achieved 98.80% training accuracy and 98.33% validation accuracy. Confusion matrix testing showed good performance, with an accuracy of 98% and for the species Oncomelania hupensis lindoensis had a recall of 93%, precision of 100%, F1-score of 97%, and the resulting AUC value of 99.7%. This research successfully developed an efficient identification system, which is expected to help health surveillance personnel in accelerating the identification process of schistosomiasis intermediate hosts.Schistosomiasis caused by the Schistosoma japonicum worm is a significant health problem in Indonesia, especially in endemic areas such as the Napu Plateau and Bada Plateau. The main problem in controlling this disease is the difficulty in rapid and accurate identification of Oncomelania hupensis lindoensis snails as intermediate hosts of the parasite. This research aims to develop an artificial intelligence-based system that can efficiently identify the snail species. The stages of this research include collecting snail image data from the Central Sulawesi Provincial Health Office, consisting of 2100 images covering seven snail species, then processed through preprocessing and augmentation stages. The model applied was EfficientNet-B0. The results showed that the EfficientNet-B0 model achieved 98.80% training accuracy and 98.33% validation accuracy. Confusion matrix testing showed good performance, with an accuracy of 98% and for the species Oncomelania hupensis lindoensis had a recall of 93%, precision of 100%, F1-score of 97%, and the resulting AUC value of 99.7%. This research successfully developed an efficient identification system, which is expected to help health surveillance personnel in accelerating the identification process of schistosomiasis intermediate hosts
Evaluation of Scalability and Resilience of Hyperledger Fabric in Blockchain Implementation for Diploma Management
No full textThis research aims to evaluate the performance of a Hyperledger Fabric-based blockchain system implemented for digital diploma management. The system is tested using the Caliper benchmark tool in various network and scalability scenarios, including normal conditions (baseline), network delay of 50ms, 100ms, 200ms, and 500ms; packet loss of 1%, 5%, 10%, and 15%; bandwidth limitation of 5 Mbps; high transaction load (scalability standard and scalability optimized); and extreme conditions in the form of Byzantine attacks with malicious nodes of 10%, 30%, and 50%. The evaluation was conducted using four key metrics: transaction success rate, failure rate, average transaction latency, and throughput (TPS). The system recorded high performance under normal network conditions with a success rate of 99.8%, latency of 0.89 seconds, and throughput of 9.7 TPS. Network disruptions such as delay, packet loss, and bandwidth limitation had only a minor impact, with the success rate remaining above 99% and latency gradually increasing. High load in the scalability scenario caused latency to increase to 27.21 seconds and failure rate to rise, but improved significantly after chaincode optimization. Meanwhile, the Byzantine scenario showed a drastic drop in performance with the success rate decreasing to 12.83% and the failure rate increasing to 87.17%. These results show that the Hyperledger Fabric-based digital diploma management system is resilient to common network disruptions and reliable at medium scale, but still requires strengthening the consensus mechanism to deal with extreme conditions and maintain reliability in environments that are not fully trusted
Optimizing Customer Data Security in Water Meter Data Management Based on RESTful API and Data Encryption Using AES-256 Algorithm
No full textGood, accurate and secure data management is certainly one of the main needs for companies that provide public services. This research aims to develop a web application-based information system to manage customer water meter data at a regional water company in Semarang. This system was built using the RESTful API architecture using the PHP programming language framework, namely Laravel and the development of web page displays using the Javascripts framework. The data used is the original database managed by the company every month which is managed using a database management system by meter reader officers. To increase the security of customer data, a cryptographic algorithm is used, namely the Advanced Encryption Standard (AES) algorithm with a 256-bit key length to secure data that is considered sensitive and contains high privacy. This system is intended for meter readers to update customer water meter data per month in an efficient and structured manner. This research uses a Research and Development (R&D) based software development method with system testing using black-box testing method to ensure application functionality and data exposure testing method to ensure data security in the database. The test results show that the system successfully manages customer water meter data in realtime per data sent and secures customer data
Comparative Analysis of CNN, Transformers, and Traditional ML for Classifying Online Gambling Spam Comments in Indonesian
No full textThe rise of user-generated content on social media and live-streaming platforms has intensified the spread of spam, particularly online gambling (Judi Online) promotions, which remain prevalent in Indonesian comment sections. This study investigates the effectiveness of various machine learning (ML) and deep learning (DL) approaches in classifying such spam content in Bahasa Indonesia. We compare five models: Support Vector Machine (SVM), Random Forest (RF), a CNN-based model, IndoBERT, and a custom lightweight transformer model named Wordformer. While IndoBERT achieves the highest performance across all metrics, it comes with high computational demands. Wordformer, in contrast, delivers a strong balance between accuracy and efficiency, outperforming traditional models while being significantly more lightweight than IndoBERT. Wordformer achieved 0.9975 accuracy and macro F1-score, surpassing SVM (0.9578) and Random Forest (0.9729), while maintaining a significantly smaller model size and fewer multiply-add operations. An extensive ablation study further explores the architectural and training design choices that influence Wordformer’s performance. The findings suggest that lightweight transformer models can offer practical, scalable solutions for spam detection in low-resource language settings without the need for large pretrained backbones