基于集成算法的青岛市二手房房价分析与预测

Analysis and Prediction of Second-Hand House Price in Qingdao Based on Integrated Algorithm

房地产业关乎国计民生,而二手房交易作为房地产市场的重要组成部分,需要长期、稳定、健康发展。二手房交易过程复杂,这对购房者而言,了解二手房价格显得尤为迫切,同时,二手房价格也是市场监管部门的关注重点。本文利用网络爬虫技术获得“链家”平台2021年度青岛市所有已成交二手房源的相关数据,进行数据预处理后,对比Lasso、随机森林、LightGBM、XGBoost四种模型的预测结果,发现XGBoost模型具有较好的预测优势。由于单一模型的局限性,本文采用Stacking算法进行模型融合,搭建RF-LG-XG模型,预测结果表明本文提出模型的预测效果优于以上单一模型。本文构建二手房价格预测模型为购房者提供了更透明、更准确的参考价格,同时为政府调整政策提供参考,促进房地产市场稳定持续发展。

The real estate industry is related to the national economy and the people’s livelihood. As an im-portant part of the real estate market, second-hand housing transactions need long-term, stable and healthy development. The transaction process of second-hand houses is complex, which is par-ticularly urgent for buyers to understand the price of second-hand houses. At the same time, the price of second-hand houses is also the focus of the market supervision department. This paper uses the web crawler technology to obtain the relevant data of all the second-hand houses that have been sold in Qingdao in 2021. Comparing the prediction results of Lasso, random forest, XGBoost and LightGBM, it is found that XGBoost model has better prediction advantages. Due to the limitations of a single model, this paper uses Stacking algorithm for model fusion, and the prediction effect of the RF-LG-XG model is better than the above single model. This paper constructs a second-hand housing price prediction model to provide more transparent and accurate reference prices for home buyers, as well as reference for the government to adjust policies, which is of great practical significance.