<![CDATA[Abstrak - Pemantauan kondisi perairan secara berkelanjutan merupakan aspek penting dalam mendukung pengelolaan sumber daya air dan mitigasi potensi bencana. Penelitian ini bertujuan untuk merancang, mengimplementasikan, dan menguji sistem Smart Water Monitoring berbasis Internet of Things (IoT) dan Machine Learning untuk memantau parameter ketinggian air, gelombang, dan suhu air secara real-time. Sistem dikembangkan menggunakan sensor ultrasonik dan sensor suhu yang terintegrasi dengan mikrokontroler serta dikoneksikan ke platform berbasis web untuk visualisasi data. Data hasil pengukuran dikirimkan melalui jaringan internet dan disimpan dalam basis data sebagai bahan analisis lanjutan. Metode Machine Learning diterapkan untuk menganalisis pola data dan mendeteksi kondisi anomali berdasarkan perubahan parameter air yang signifikan. Pengujian sistem menunjukkan bahwa perangkat IoT mampu melakukan akuisisi dan transmisi data secara stabil, sementara model Machine Learning yang digunakan memberikan performa yang baik dalam mengidentifikasi kondisi tidak normal pada data perairan. Hasil penelitian ini menunjukkan bahwa integrasi IoT dan Machine Learning dapat menjadi solusi yang efektif dan efisien untuk sistem pemantauan kondisi air secara cerdas dan berkelanjutan.Kata kunci: Sistem Logging; Otentikasi Dua Faktor; Rate Limiter; Machine Learning; Deteksi Anomali; Abstract - The development of modern cyber threats requires network security systems to have adaptive and integrated detection capabilities. This research aims to develop and test a prototype web-based network logging system equipped with a multi-layered authentication mechanism and anomaly pattern analysis using Machine Learning (ML). The system was developed using the Flask (Python) framework and tested online. The system's security components include Google reCAPTCHA and Two-Factor Authentication (OTP) for access protection, as well as the implementation of a Rate Limiter to mitigate low-rate distributed (multi-IP) attacks. The collected activity log data was then used to train two classification models, namely Decision Tree and Random Forest, with the main feature being the frequency of activity per IP within 60 seconds. Test results show that the Rate Limiter system successfully limits low-volume attacks. Meanwhile, ML performance analysis proves the effectiveness of the proposed method, where Decision Tree achieves perfect accuracy of 100.0% and an F1-Score of 1.0 in classifying anomalous activities in structured log datasets. This implementation demonstrates that the integration of secure logging with Machine Learning provides a strong foundation for the development of intelligent and efficient real-time threat detection systems.Keywords: Logging System; Two-Factor Authentication; Rate Limiter; Machine Learning; Anomaly Detection;]]>
Copyrights © 2026
