Comparative study of boston house price prediction models based on gan data augmentation

Abstract

This study aims to compare the performance of four machine learning models in the Boston housing price prediction task, and explore whether GAN data augmentation can improve the prediction performance of the model. Experimental results indicate that on the original data set, The CatBoost model demonstrated superior predictive performance, with an MSE of 4.76, RMSE of 2.18, MAE of 1.65, and R² of 0.94. However, no significant performance improvement was observed when GAN data augmentation was applied to these models.