24.04.169
006.31 - Machine Learning
Karya Ilmiah - Skripsi (S1) - Reference
Machine Learning, Operation Systems,
<p>Detecting data anomalies in the operational<br /> process of oil and gas pipelines is very important to reduce the<br /> risk of disasters, which can adversely affect human safety, the<br /> environment and financial aspects. Failure to do so can lead to<br /> catastrophic results. The problem is also supported by several<br /> catastrophic events that have occurred in several areas of oil and<br /> gas production facilities in several regions. To solve this<br /> problem, it is necessary to implement a suitable monitoring<br /> system that aims to prevent potential losses caused by leaks or<br /> over-pressurization of natural gas pipelines. Among the many<br /> machine learning algorithms available for anomaly detection<br /> such as Feed Forward Neural Network, Linear Regression,<br /> KNN, Random Forest, and Support Vector Machine and<br /> unsupervised machine learning models such as Principal<br /> Component Analysis (PCA) and Hierarchical clustering, One-<br /> Class SVM and Isolation Forest are the most prominent.<br /> However, these algorithms have their own advantages and<br /> disadvantages regarding their performance. This study aims to<br /> compare the performance of machine learning algorithms in<br /> classifying and detecting data anomalies in offshore natural gas<br /> pipeline operational datasets. The assessment is based on ROC-<br /> AUC Curve, Confusion Matrix, Sensitivity, and Specificity. The<br /> findings indicate that the Isolation Forest model outperforms<br /> the One-Class SVM, with a ROC-AUC value of 90%, compared<br /> to the One-Class SVM's value of only 61%. Furthermore, the<br /> Isolation Forest exhibits a Sensitivity value of 98%, in contrast<br /> to the One-Class SVM's 41%, and a Specificity of 81%,<br /> compared to the One-Class SVM's 80%.</p>
Tersedia 1 dari total 1 Koleksi
Nama | KELVYN LUKITO |
Jenis | Perorangan |
Penyunting | Hasmawati, Aditya Firman Ihsan |
Penerjemah |
Nama | Universitas Telkom, S1 Informatika |
Kota | Bandung |
Tahun | 2024 |
Harga sewa | IDR 0,00 |
Denda harian | IDR 0,00 |
Jenis | Non-Sirkulasi |