Research on Price Evaluation of Second-hand Container Ship based on GA-BP Neural Network

Abstract

As a capital investment, whether the investment in second-hand container ships can bring benefits to the enterprise depends on whether the benefits of ship operation can repay the investment and obtain certain profits. The purchase price of second-hand ships is the key factor, so an evaluation model is built to predict the price of second-hand container ships and evaluate its rationality. To help enterprises better understand the actual price of ships when buying and selling ships, and hope to improve the market evaluation system. In order to accurately evaluate the price of second-hand ships, the Clarkson database was used to obtain 362 container ship transaction data from December 2019 to December 2022, and a GA-BP network neural model based on ship age, deadweight tonnage and new shipbuilding price was established. The weights and thresholds were optimized through genetic algorithms, and then assigned to the BP neural model. The test results show that the mean square error is 0.0185, the evaluation error is basically 20%, and the error within 10% accounts for nearly half. It is proved that the model can preliminarily evaluate the transaction price of second-hand ships, and has certain feasibility and rationality.