Analisis Relasi Antara Risk Event dan Risk Agent Pada HORshe Menggunakan Machine Learning : Studi Kasus Laporan Investigasi Kecelakaan KNKT

Putra, Kelvin Ardian Syah (2025) Analisis Relasi Antara Risk Event dan Risk Agent Pada HORshe Menggunakan Machine Learning : Studi Kasus Laporan Investigasi Kecelakaan KNKT. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5010211035-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (5MB) | Request a copy

Abstract

Peningkatan jumlah pengguna transportasi darat juga menyebabkan jumlah kecelakaan lalu lintas semakin meningkat. Perlu adanya identifikasi terhadap risk event dan risk agent dari suatu kecelakaan yang terdapat didalam laporan investigasi kecelakaan milik KNKT untuk upaya mitigasi dan pengurangan tingkat kecelakaan. Namun hingga saat ini penilaian terhadap relasi antara risk agent dan risk event dalam metode HORshe untuk memitigasi suatu risiko kecelakaan masih bersifat subjektif dan tradisional sehingga diperlukan improvement dalam penilaian tersebut salah satunya menggunakan machine learning. Penelitian ini memiliki tujuan yaitu mengimplementasikan Machine learning terutama teknik embedding model untuk menganalisis nilai relasi antara risk event dan risk agent dalam metode HORshe berdasarkan konteks kejadian. Didapatkan total 2019 data risk event, risk agent, konteks kejadian, serta nilai relasi / keterkaitan antara risk event dan risk agent yang berasal dari laporan investigasi KNKT tahun 2014 hingga 2024. Model machine learning yang diguanakan dalam penelitian ini adalah BERT dengan arsitektur simple transformers dan model bahasa indobenchmark/indobert-base-p2 dikarenakan model ini memiliki nilai f1-score di kelas penting yaitu nilai relasi 3 dan 9 lebih tinggi daripada yang lain yaitu 0.8054 dan 0. 8225 serta memiliki akurasi paling tinggi yaitu 0.8457 dibandingkan dengan model lainnya yaitu TF-IDF dan Sentence-BERT dengan model bahasa indobenchmark/indobert-large-p2 dan wira-pratama/k3mbed-tsdae-v1. Machine Learning yang dibuat dapat meningkatkan performansi, akurasi, dan juga efisiensi dari proses analisis nilai relasi risk event dan risk agent sebagai langkah untuk mitigasi risiko yang dulunya masih dilakukan secara manual.
=============================================================================================================================================
The increase in the number of land transportation users has also caused the number of traffic accidents to rise. There is a need to identify the risk events and risk agents of an accident found in the KNKT accident investigation report for the purpose of mitigation and reduction of accident rates. However, until now, the assessment of the relationship between risk agents and risk events in the HORshe method for mitigating accident risks remains subjective and traditional, necessitating improvements in this assessment, one of which is through the use of machine learning. This research aims to implement Machine Learning, particularly embedding model techniques, to analyze the relationship value between risk events and risk agents in the HORshe method based on the context of incidents. A total of 2019 data points on risk events, risk agents, incident contexts, and the relational values/associations between risk events and risk agents were obtained from KNKT investigation reports from 2014 to 2024. The machine learning model used in this study is BERT with a simple transformers architecture and the indobenchmark/indobert-base-p2 language model because this model has higher f1-scores in the important classes, namely relation scores 3 and 9, at 0.8054 and 0.8225, respectively, and the highest accuracy at 0.8457 compared to other models such as TF-IDF and Sentence-BERT with the indobenchmark/indobert-large-p2 and wira-pratama/k3mbed-tsdae-v1 language models. The Machine Learning developed can improve the performance, accuracy, and efficiency of the analysis process of risk event and risk agent relationship values as a step for risk mitigation that was previously done manually.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	HORshe, Kecelakaan, Machine Learning, Risk Event, Risk Agent. HORshe, Accident, Machine Learning, Risk Event, Risk Agent.
Subjects:	T Technology > T Technology (General) > T174.5 Technology--Risk assessment. T Technology > T Technology (General) > T55 Industrial Safety
Divisions:	Faculty of Industrial Technology and Systems Engineering (INDSYS) > Industrial Engineering > 26201-(S1) Undergraduate Thesis
Depositing User:	Kelvin Ardian Syah Putra
Date Deposited:	01 Jul 2025 04:32
Last Modified:	01 Jul 2025 04:32
URI:	http://repository.its.ac.id/id/eprint/119299

Actions (login required)

View Item