Application and Optimization of Machine Learning-Based Ensemble Learning Methods in Sailboat Valuation

With the growing sailboat market, accurate valuation of sailboats is crucial for both buyers and sellers. This study proposes an ensemble learning method based on machine learning to enhance the accuracy and efficiency of used sailboat valuation. Through an in-depth analysis of the sailboat dataset, we first conducted data preprocessing, including cleaning missing values and outliers, as well as constructing a database and quantifying indicators. Building on this foundation, we employed three different machine learning algorithms Random Forest, BP Neural Network, and XGBOOST to build individual predictive models. To further improve predictive performance, we adopted an ensemble learning strategy, training a linear regression model as a secondary learner using the predictive results from the aforementioned models. This Stacking approach not only combines the strengths of each base learner but also enhances the overall model's generalization capability by optimizing hyperparameters. Experimental results demonstrate that the integrated and optimized model excels in sailboat valuation tasks, achieving higher accuracy on both training and test sets compared to individual models. Additionally, we explored the impact of regional economic factors on sailboat prices, finding a significant correlation between economic indicators of different regions and sailboat prices.