Integrating GBDT regression and factor analysis for accurate prediction of used sailboat prices

Abstract

Similar to many luxury goods, the value of sailboats fluctuates with the age of the boat and market conditions. To comprehensively understand the sailboat market, our team utilized web crawlers to collect data on second-hand sailboats and relevant regional factors (economic, geographic, tourism). We established a unified prediction model that integrates sailboat characteristics and regional influences. Using a factor analysis model, combined with a Gradient Boosting Decision Tree (GBDT) regression model, we evaluated the regional impact on second-hand sailboat prices. The combined model demonstrated high accuracy in predicting the prices of monohull and catamaran sailboats. The study further explored the applicability of the model to the Hong Kong market and validated its effectiveness with different datasets. The results indicated substantial demand and tremendous market potential in the Hong Kong market. Finally, through ARIMA time series forecasting, statistical histograms, and correlation analysis, the study revealed relationships between year and price, regional disparities, and key price-influencing factors.