Research on House Price Forecast Based on Hyper Parameter Optimization Gradient Boosting Regression Model

In order to study the prediction accuracy of different regression models, this paper selects housing price prediction as a typical research object, and proposes a hyper parameter optimization gradient boosting regression model (HPOGBR). Firstly, the data is preprocessed and the key features are extracted through visualization and basic model pretraining. Then, multiple machine learning regression models are used for comparison with our HPOGBR model base on evaluation indicators. Finally, the optimal model and parameters are selected by grid search. The experimental results show that our proposed HPOGBR model has better accuracy and convergence over existing machine learning regression models.