MLP模型在房地产批量评估中的应用研究

Application of Multi-layer Perceptron Model in Real Estate Mass Appraisal

第十三届全国人大常务委员会第三十一次会议通过《授权国务院在部分地区开展房地产税改革试点工作的决定(草案)》,这意味着房地产税改革进入实质性推进阶段。在此背景下,以批量评估为基础的房地产价值评估为房地产税的开征提供了技术支撑。本文运用人工神经网络中多层感知机(简称MLP)模型,以石家庄市长安区二手住宅为样本,进行实证研究,并将其结果与特征价格模型所得结果进行对比;同时,对裕华区二手住宅做以同样的实证分析。最终得出:相比经典特征价格模型,运用多层感知机模型得出的评估结果与真实交易值的误差更小,匹配度更优。

The 31st meeting of the Standing Committee of the 13th National People’s Congress adopted the decision(Draft) of authorizing the State Council to carry out the pilot work of real estate tax reform in some regions on October 23, 2021. This means that the real estate tax reform has entered a substantive stage.In this context, the real estate value evaluation, which based on mass appraisal,provides technical support for the collection of real estate tax.This paper uses the multi-layer perceptron(MLP) model of artificial neural network in making an empirical study,which takes the second-hand housing in Chang’an District of Shijiazhuang as a sample, and compares the result with the result of Hedonic price model and makes the same empirical analysis on the second-hand housing in Yu hua District.Finally, compared with the classical Hedonic price model, the error between the evaluation results obtained by using the multi-layer perceptron model and the real transaction value is smaller and the matching degree is better.