Risk Factors Analysis Modeling for Ship Collision Accident in Inland River Based on Text Mining

Abstract

Inland waterway ship collision is a typical high-risk maritime accident which caused by many factors. In order to clarify the causal factors of collision of inland watercraft, the risk of inland river collision is analyzed. First of all, 419 ship collision accidents in the Yangtze River inland waterway from 2013 to 2017 are selected as the ext mining corpus. The human factors, ship factors, natural environment factors and management factors in the corpus database, will be applied as the target base. R language and text mining method is utilized to get a high dimensional and sparse original feature vector space set, which contains the features and the weight of the features. The chi-squared statistic is used to reduce the dimensionality of the set and 33 dimension features set which representing the ship collision risk factors are generated. Finally, the word cloud package in R language is invoked to make the date visualization, which provide a theoretical basis for the pre-control of inland waterway ship collision.