Model Prediction Study of Boston Dataset Based on Random Forest Fusion Algorithm

This study is based on the random forest fusion algorithm for model prediction of the Boston dataset, aiming to explore the application of statistics in biomedical research. Through data visualization and a variety of statistical methods, this study provides an in-depth analysis of the characteristics of the dataset and variable relationships. The study first introduces the importance of statistics in the biomedical field, including the application of descriptive statistics, inferential statistics, Bayesian statistics, probability theory, regression analysis, multivariate analysis, and survival analysis. Subsequently, this study elaborates on the Random Forest algorithm and constructs a hybrid model to improve the prediction accuracy by fusing the Gradient Boosted Tree (GBDT) model. Experimental results show that the fusion model performs well in reducing prediction errors and improving model stability, especially in the house price prediction task, where the fusion model outperforms the single model in terms of mean square error (MSE) and coefficient of determination (R²). By analyzing the Boston house price dataset, this study finds that variables such as the average number of rooms per dwelling (RM), the percentage of low-income people (LSTAT), and air quality (NOX) have particularly significant effects on house prices. This study not only verifies the validity of the fusion model, but also provides a scientific basis for urban planning and policy making. Future research can further optimize the model, combine deep learning techniques, expand the dataset and variables, and deepen interdisciplinary applications to enhance the universality and practicality of the research results.