應用羅吉斯迴歸建立油輪重大意外事故預測模型

Forecasting Probability of Tanker Accidents Using Logistic Regression Model

本研究應用羅吉斯迴歸模型，調查嚴重之油輪事故原因，統計數據來自1980至2010年期間國際海事組織的全球綜合航運信息系統。研究主要目的在辨識導致油輪全損的事故之實質危險因素分析，發現油輪類型和事故的類型具有顯著與全損嚴重事故相關，特別是小型油輪意外事故之沉沒事件，將導致船東高幅度資產的損失危險。因此小型油輪安全管理應著重於降低意外事故所造成沉沒之發生頻率；具體來說，採取安全管理策略，應以改善機械設備維護及注重人員安全培訓為主。 This study applied the logistic regression model to investigate the causes of serious oil tanker accidents. The statistical data was derived from the IMO's global integrated shipping information system from 1980 to 2010. The purpose of the study was to identify the actual risk factors for accidents that caused the total loss of the tanker. It was found that the type of the tanker and the type of accident were significantly related to the total accident loss. In particular, the sinking of a small tanker accident would lead to high risk of loss of assets. Therefore, the safety management of small oil tankers should focus on reducing the frequency of sinking caused by accidents; specifically, safety management strategies should be adopted to improve the maintenance of machinery and equipment; focus on personnel safety training.