Prediction Model of Ship Anchoring in Submarine Cable Protection Area Based on Big Data Platform

The present study aimed to investigate the problem of frequent submarine cable failures caused by ship anchoring in the submarine cable are using an anchoring prediction model based on a big data platform. The model used the speed change rate of the ship as the feature vector, the density-based spatial clustering of applications with noise(DBSCAN) density clustering algorithm for outlier detection, and replaced the outliers with the average. The hash bucket strategy resolved the issue of inconsistent data intervals and used logistic regression and nonlinearity. The four algorithms of logistic regression, support vector machine (SVM), and non-linear SVM were trained, and the accuracy and recall rate of the model was about 90%, which could predict the ship's anchoring behavior adequately. Using this model, the artificial intelligence system can predict the anchoring behavior of ships accurately, prohibit ships from anchoring in submarine cable protection areas in advance, and reduce the risk of submarine cables being anchored. It has a good effect on improving the efficiency of sea area management and reducing the cost of anchoring accidents.