인공신경망을 활용한 중고선가 예측모형 개발에 관한 연구

초록

Purpose : The purpose of this study is to develop forecasting model for secondhand drybulk ships using artificial neural network (ANN) model. Research design, data and methodology : Price of capesize bulk ship is target variable since ship price is relatively high with more variability. Multivariate linear regression (LR) and ANN are adopted with market model, index model and market-index model. Forecasting errors are compared to select the best model with better accuracy. Results : ANN outperforms LR analysis in general. It is found that, among the three models, market-index model with ANN is the most accurate. Conclusions : This study is meaningful in that it applies regression model and deep learning model to secondhand ship price forecasting. Public agencies and consulting firms in maritime sector are able to adopt findings from this research.

키워드