Correlation Analysis Of Marine Casulties Based On The Improved Apirori Algorithm

Marine accidents have always been one of the most important issues in the shipping field.The analysis of accident information is an important way to identify maritime risks.Based on the database of maritime accidents,thesis systematically reviews the development process of research methods for maritime accidents.It also proposes the development of computer information processing technology and the era of big data.Data mining is used to conduct relevant research on maritime accident information databases.In view of the current statistics format and description of unmanned maritime accident investigation,taking statistics database of Zhejiang Maritime Safety Administration(MSA)as an example,the original database is combed and cleaned,and a standardized ship traffic accident information database is established.And transforms and encodes the original statistical information and description of the traffic accident on the ship.At the same time,it uses the method of mathematical statistics to identify the proportions of all causes caused by human,ship,cargo,environment and management.Causal factors of traffic accidents and degree of influence.In order to further studies on the correlation among various information about ship traffic accidents,a combination mining algorithm model of k-medoids and Apriori is constructed to deeply mine the relevant information in the standard database.Eight correlation rules of collision accidents were extracted under the condition of 20%of the support threshold and 50%of the confidence threshold.The association of 12 non-collision accidents was extracted under the condition of 10%of the support threshold and 50%of the confidence threshold rule.The experimental results show that both k-medoids and Apriori based mining algorithms are superior to the traditional Apriori algorithm in terms of both the promotion value and the mining accuracy.Finally,by analyzing the association rules,the paper quantitatively analyzes the relationship among the various marine accident information,identifies the characteristics of maritime traffic risk in the Zhejiang waters,and from the three levels providing measures and suggestions to maritime authorities,shipping companies and officer on watch(OOW).The results provide maritime authorities with theoretical support for maritime surveillance,risk investigation and identification of risk sources.They also providing shipping companies with advice on vessel and crew management and providing crew with references decision-making to identify the complex maritime environment for preventing the risks.