中国房价构成与预测的仿真分析  

Simulation Analysis of China's Housing Prices Constitution and Prediction

作　　者：陆丽丽[1] 胡斌[1] 李辉[1] 端木怡婷[2] 

机构地区：[1]解放军理工大学指挥信息系统学院 [2]中国人民解放军95851部队

出　　处：《计算机仿真》2014年第3期230-238,共9页Computer Simulation

摘　　要：研究房价准确预测问题,结合近年来国内房价易涨难跌、难以调控的问题,提出利用回归分析和BP神经网络的相关知识,建立了房价构成与预测模型。首先,分析房价构成因素,通过多元线性回归分析方法建立房价构成模型,并通过仿真得到了影响房价的主要因素,在此基础上,利用BP神经网络构建房价预测模型;根据历史统计数据分别预测07、08、09连续三年的房价,并将其与实际值进行比对验证仿真模型的可靠性及有效性。最后,结合2009年的数据参数,分别分析各个主要因素如何对房价产生影响,仿真结果表明,为房价的准确预测提供了依据。Recently, domestic housing prices feel hard to down in a short term and is difficult to regulate. Consid- ering that, this paper constructed a constitution and prediction model of housing price, basing on Regression Analysis and Back Propagation Network. First of all, we analyzed the factors that influence housing prices and then used Mul- tiple Regression Analysis method to construct a constitution model. From that model, we can achieve the most essen- tial factors of all. On the basis of those factors, we used the knowledge of BP Network to construct a prediction mod- el. Using that model, we predicted the housing prices of year 2007, 2008, 2009 and validated the model by compa- ring them with the practical data. At last, we studied how the most essential factors influence the housing price sepa- rately, trying to dig out the underlying reasons behind the incredible price.

关 键 词：房价构成与预测模型 回归分析 神经网络 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]