Research on Ship Target Classifications and Recognitions by Use of Support Vector Machine of Machine Learning

Abstract

This research applies the machine learning models to the classification and identification in ship types with the acoustic data of noise By importing the simulation data of three different ship types of noise with different signal noise ratio into the support vector machine model of the machine learning theories including multiple classes classification model Naive Bayes classify model and Error-correcting output codes model we are able to classify and predict the ship types we expect the results to be more precisely with the improvement on the ship acoustic data of noise processing in the future The results show that Error-correcting output codes model presents better in the prediction of ship types than other two models with the data of characteristic frequency against speed Also by the increasing of signal noise ratio level it presents better for characteristic frequency against different constant Key words: Support Vector Machine Ship Radiation Sound Field Machine Learning

Date of Award	2020
Original language	English
Supervisor	Gee-Pinn James Too (Supervisor)