Bootstrap-Tobit model for maritime accident economic loss considering underreporting issues. Issue 4 (10th December 2021)

• Record Type: Journal Article
• Authors: Li, Guorong; Weng, Jinxian; Fu, Shanshan
• Is Part Of: Transportmetrica. Volume 17:Issue 4(2021)
• Page Start: 1055
• Page End: 1076
• Publication Date: 2021-12-10
• Subjects: Maritime accident -- bootstrap sampling -- underreporting data -- reliability; Transportation -- Periodicals; Transportation -- Research -- Periodicals; 388.072
• DOI: 10.1080/23249935.2020.1829169
• Languages: English
• ISSNs: 2324-9935

Abstract

Abstract : The underreporting of maritime accident data is a major concern in maritime transportation safety analysis. In order to avoid biased effects caused by the underreporting issue, this study aims to propose a Bootstrap-Tobit regression model for the estimation of economic loss resulting from maritime accidents. The questionnaire survey method is employed to determine the possible underreporting rates of maritime accident data and their occurrence probabilities, while the bootstrap sampling method is adopted to supplement the underreported data by generating a number of bootstrap samples. With ten year's archived maritime accident data, a case study is created to calibrate the proposed model. The proposed Bootstrap-Tobit model has captured the impacts of underreporting issues on the marginal effects of the economic loss contributory factors. The proposed model could supplement sufficient underreported data so that the results are more reliable and convincible.