Fitria, Annisa Nurul (2025) Prediksi Cuaca Maritim Menggunakan Dynamic Bayesian Network. Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 2043211036-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
2043211036-Undergraduate_Thesis.pdf Restricted to Repository staff only until 1 April 2027. Download (1MB) | Request a copy |
Abstract
Indonesia merupakan negara kepulauan terbesar di dunia, sehingga menjadikan transportasi laut sebagai sarana vital untuk pergerakan barang dan orang. Kapal sebagai moda transportasi laut memainkan peran yang sangat penting dalam mendukung perdagangan, baik di dalam negeri maupun luar negeri. Tingginya penggunaan transportasi laut dapat meningkatkan terjadinya kecelakaan, terutama yang disebabkan oleh cuaca buruk. Cuaca maritim seperti gelombang tinggi dan angin kencang berdampak buruk terhadap keselamatan pelayaran dan sering kali menyebabkan kecelakaan serius, sehingga kondisi cuaca maritim sangat penting terhadap semua aspek operasional dan keselamatan pelayaran. Prakiraan cuaca maritim dapat membantu menghindari kejadian buruk di laut, tetapi cuaca sulit diprediksi terutama di wilayah perairan sehingga diperlukan penelitian lebih lanjut terkait prediksi cuaca maritim. Penelitian ini menggunakan metode Dynamic Bayesian Network (DBN) untuk memprediksi cuaca maritim, sehingga dapat mengidentifikasi faktor-faktor utama yang memengaruhi cuaca maritim dengan menangkap ketergantungan temporal di Indonesia. Model prediksi cuaca maritim di Indonesia yang didapatkan melalui implementasi Dynamic Bayesian Network menghasilkan nilai akurasi sebesar 97,61%. Nilai Area Under the Curve (AUC) untuk kategori Significant Wave Height of Combined Wind Waves and Swell kategori "Tenang", “Rendah”, “Sedang”, dan “Tinggi” secara berurut memiliki nilai sebesar 0,904292, 0,974501, 0,968113, 0,985384. Nilai ini mendekati 1 mengindikasikan bahwa model Dynamic Bayesian Network memiliki kemampuan yang sangat baik dalam membedakan masing-masing kelas.
===================================================================================================================================
Indonesia is the largest archipelago in the world, making sea transportation a vital means for the movement of goods and people. Ships as a mode of marine transportation play a very important role in supporting trade, both domestically and abroad. The high use of marine transportation can increase the occurrence of accidents, especially those caused by bad weather. Maritime weather such as high waves and strong winds adversely affect shipping safety and often cause serious accidents, so maritime weather conditions are critical to all aspects of shipping operations and safety. Maritime weather forecasts can help avoid adverse events at sea, but weather is difficult to predict especially in water areas so further research is needed regarding maritime weather prediction. This research uses the Dynamic Bayesian Network (DBN) method to predict maritime weather, so as to identify the main factors that affect maritime weather by capturing temporal dependencies in Indonesia. The maritime weather prediction model in Indonesia obtained through the implementation of Dynamic Bayesian Network produces an accuracy value of 97.61%. The Area Under the Curve (AUC) value for the Significant Wave Height of Combined Wind Waves and Swell category of “Calm”, “Low”, “Medium”, and “High” respectively has a value of 0.904292, 0.974501, 0.968113, 0.985384. This value is close to 1 indicating that the Dynamic Bayesian Network model has a very good ability to distinguish each class.
Actions (login required)
 |
View Item |