Year: 2025
Type: article
Abstract: House is a basic demand for living. It is directly related to people's happiness. Affected by many factors such as the economy condition and the policy, house prices are always dynamic. House price prediction can provide decision support for investors, so it is important to optimize the prediction model. Many machine learning methods are applied to house price prediction, such as support vector machine and random forest. This paper predicts the house price of Boston. To solve the multicollinearity problem in the dataset, this paper builds a BP neural network model using Bayesian regularization. The fitting analysis of the data samples shows the total R-value is 0.9824. The loss function converges to the global minimum within 100 epochs. The experiment indicates that the model is both efficient and reliable in forecasting house prices. This research provides a method to deal with the multicollinearity problem in the dataset, which improves the generalization ability and predictive accuracy of the model.
DOI: 10.54097/2vq1wq81
Author: Zuxin Chen
Institution: Shenzhen University
Language: English
Topic: Housing Market and Economics
Subfield: Economics and Econometrics
Domain: Social Sciences
Open Access status: diamond
OpenAlex ID: W4407728143