Pemodelan Prediksi Tingkat Keparahan Kecelakaan Kapal Menggunakan Bayesian Network

Prasetyowati, Selviana (2024) Pemodelan Prediksi Tingkat Keparahan Kecelakaan Kapal Menggunakan Bayesian Network. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2043211068_Undergraduate_Thesis.pdf Restricted to Repository staff only until 1 April 2027. Download (4MB) | Request a copy

Abstract

Indonesia memiliki wilayah perairan seluas 5,95 juta km², menjadikan pelayaran sebagai tulang punggung transportasi dan ekonomi nasional. Meski demikian, keselamatan pelayaran masih menjadi perhatian utama dan kecelakaan tersebut dikategorikan sebagai Very Serious Marine Casualty (VSMC) atau Marine Incident (MI) sesuai tingkat keparahannya. Penelitian ini bertujuan mengembangkan model prediksi tingkat keparahan kecelakaan kapal menggunakan Bayesian Network (BN), yang memetakan hubungan sebab-akibat antar variabel risiko. Objek penelitian mencakup 111 kecelakaan kapal penyeberangan di Indonesia pada 2007-2023, yang bersumber dari investigasi KNKT dan Mahkamah Pelayaran. Sesuai Undang-Undang RI No. 17 Tahun 2008, kecelakaan dikelompokkan dalam tiga kategori, yakni regulator, operator, dan fasilitator. Data dibagi menggunakan metode stratified holdout untuk menjaga proporsi kelas antara VSMC dan MI. Berdasarkan hasil analisis statistika deskriptif, Crew Deck adalah unit yang memiliki jumlah keterlibatan tertinggi diantara unit lain pada kelompok operasional dan teknis kapal. Hasil analisis diagram tornado menunjukkan keterlibatan Cargo Owner meningkatkan peluang terjadinya VSMC hingga 0,837, sedangkan ketidakterlibatan menurunkannya menjadi 0,385. Nilai Strong of Influence membuktikan bahwa Cargo Owner memiliki peluang 0,909 untuk menjadikan unit lain seperti Expedition Company terlibat dalam kecelakaan kapal. Evaluasi model menunjukkan akurasi 81,82%, kemampuan model memprediksi VSMC 80%, dan kemampuan model memprediksi MI 83,33%. Penelitian ini memberikan wawasan baru bagi otoritas maritim dan operator kapal dalam mengantisipasi kecelakaan dengan mengoptimalkan kinerja unit terkait dan merekomendasikan perbaikan Standar Operasional Prosedur (SOP) keselamatan. Selain berkontribusi pada literatur akademik, model ini mendukung pengambilan keputusan berbasis data untuk meningkatkan keselamatan transportasi maritim Indonesia.
===================================================================================================================================
Indonesia has a water area of 5.95 million km², making shipping the backbone of transportation and the national economy. However, shipping safety is still a major concern and the accident is categorized as Very Serious Marine Casualty (VSMC) or Marine Incident (MI) according to its severity. This study aims to develop a prediction model for the severity of ship accidents using the Bayesian Network (BN), which maps the cause-and-effect relationship between risk variables. The object of the study includes 111 ferry accidents in Indonesia in 2007-2023, which are sourced from the investigation of the KNKT and the Shipping Court. In accordance with Indonesian Law No. 17 of 2008, accidents are grouped into three categories, namely regulators, operators, and facilitators. The data were divided using the stratified holdout method to maintain the class proportion between VSMC and MI. Based on the results of descriptive statistical analysis, the Crew Deck is the unit that has the highest number of involvement among other units in the ship's operational and technical groups. The results of the analysis of the tornado chart showed that Cargo Owner's involvement increased the chance of VSMC to 0.837, while non-involvement decreased it to 0.385. The Strong of Influence value proves that the Cargo Owner has a 0.909 chance of getting another unit like the Expedition Company involved in a ship accident. The model evaluation showed an accuracy of 81.82%, the ability of the model to predict VSMC was 80%, and the ability of the model to predict MI was 83.33%. This study provides new insights for maritime authorities and ship operators in anticipating accidents by optimizing the performance of related units and recommending improvements to safety Standard Operating Procedures (SOPs). In addition to contributing to the academic literature, this model supports data-driven decision-making to improve the safety of Indonesia's maritime transportation.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Bayesian Network, Keparahan Kecelakaan Kapal, Keselamatan Maritim, Marine Incident, dan Very Serious Marine Casualty, Bayesian Network, Ship Accident Severity, Maritime Safety, Marine Incident, and Very Serious Marine Casualty.
Subjects:	Q Science > QA Mathematics > QA279.5 Bayesian statistical decision theory.
Divisions:	Faculty of Vocational > 49501-Business Statistics
Depositing User:	Selviana Prasetyowati
Date Deposited:	08 Apr 2025 05:31
Last Modified:	08 Apr 2025 05:33
URI:	http://repository.its.ac.id/id/eprint/119002

Actions (login required)

View Item