Analysis and Prediction of Sailboat Price Based on Improved Extreme Gradient Boosting

In order to study the price changes of sailboats, the paper establishes a prediction model for used sailboat prices based on data and analyzes regional effects. Data processing includes eliminating missing data and handling outliers by using box plots. Regression models are built using linear regression, extreme gradient boosting, and random forest analysis for monohull and catamaran sailboats. An improved extreme gradient boosting model is proposed, integrating the advantages of previous models to optimize accuracy by 0.01 % – 2.1%. The shapley additive explanation model is used to analyze the impact of region on price, revealing a regional effect influenced by GDP per capita. The model is then applied to explore second-hand sailboat trading rules in the Hong Kong market, with paired sample T -tests confirming a regional effect on prices.