Research on Second-Hand Sailboat Prices Based on Machine Learning

Abstract

With the development of the international shipping market and changes in second-hand ship prices, the operation and trade of second-hand ships are quite active. To accurately evaluate the price of second-hand ships, it is particularly important to establish a pricing model for second-hand ships. The first part of this article uses X-GBoost and PCA to rank and reduce the importance of features, and obtains four principal components. Based on 5 indicators and their importance ratios, establish an optimized X-GBoost regression model for Sine Cosine Algorithm (SCA) to predict the price of second-hand ships. In order to further investigate the impact of regional characteristics on prices, local population density, per capita GDP, and cargo throughput were selected to further classify regional characteristics. A linear regression model was established to investigate the impact of population density, per capita GDP, and cargo throughput on regional error prices, and to explore the impact of various regional characteristics on error prices. Divide sailboats into three categories – low-end sailboats, mid-range sailboats, and high-end sailboats, and discuss the impact of regional characteristics on prices. The study found that there is a significant difference in the prices of low-end sailboats and mid-range sailboats between regions, indicating consistency in regional effects. Studying the pricing of second-hand sailboats can help to correctly select trading markets, develop marketing strategies, purchase ships with higher economic priority, and promote the development and prosperity of the sailing industry.