基于神经网络的房地产市场预警系统建模与分析

发布时间：2018-05-23 08:11

本文选题：房地产 + 预警　；参考：《兰州交通大学》2013年硕士论文

【摘要】：迄今为止，中国房地产行业经历了30多年的发展历程，其发展模式实现了从计划体制到市场制度的转变，但无论房地产行业在发展中处于何种阶段和运行模式，始终离不开国家、政府的调控行为。1998年以前，国家对住宅实行福利分房的计划经济制度，房改以后，住宅作为房地产行业的主体结构走向了市场，其发展迅猛，市场运行剧烈波动，体现在价格波动巨大，投资、投机活动频繁等，直接或间接的影响着国民经济的运行和稳定，因此加强房地产市场的监管对经济社会中的各个角色尤为重要。在上述背景下，本文开展了关于房地产市场预警系统建模与分析的研究，首先阅读了大量文献，了解到国内外有关房地产市场监管和预警的现状，着重介绍了几个典型代表国家的房地产市场监管体系和国内房地产预警研究的进展，为找到合理的预警方案奠定了基础；其次，本文阐述了房地产周期理论、房地产预警理论和要素、房地产市场波动成因理论，然后将经济学原理和房地产行业相结合，为下文指标的选取等后续工作奠定了理论基础。本文最终选取的是基于神经网络的房地产市场预警手段,首先介绍了预警建模需要的准备工作，即数据预处理，包括采用时差分析筛选指标和警情警度的数值定义、区间划分；其次，选取BP神经网络算法，详细介绍了算法原理和应用算法建模的各个步骤环节，实现了BP神经网络与预警系统建模分析相结合,为实证分析的进行做了原理性论述。本文选取了天津市房地产市场为样本进行了实证分析，首先介绍了选取天津市作为样本城市和划分指标时间区间的依据，描述了天津市房地产行业的发展历程；其次，通过查询了天津市统计年鉴和天津市统计局网站，获得了天津市房地产预警指标的详细数据，确保了数据的真实性和准确性；最后，对数据进行数据分析和处理，筛选了天津市房地产市场预警指标，运用BP神经网络进行建模，借助MATLAB软件编程，，实现了BP神经网络的训练和参数的确定，并预测了2012年天津市房地产市场的警情，得出了市场运行为“热”的结论。文章最后，总结了本文的研究成果和结论，着重分析了预警系统中的不足，对未来的研究方向进行了展望，对今后预警研究的发展从制度角度提出了一些政策性建议。
[Abstract]:Up to now, China's real estate industry has experienced more than 30 years of development, its development model has realized the transformation from the planning system to the market system, but no matter what stage and operation mode the real estate industry is in the development, Before 1998, the state implemented the planned economy system of housing welfare and divided housing. After the housing reform, housing as the main structure of the real estate industry went to the market, and its development was swift and violent. The fierce fluctuations in market operation are reflected in the huge price fluctuations, frequent investment and speculative activities, which directly or indirectly affect the operation and stability of the national economy. Therefore, strengthening the supervision of the real estate market is particularly important to the economic and social roles. Under the above background, this paper has carried out the research on the modeling and analysis of the real estate market early warning system. First of all, it has read a lot of literature and learned about the current situation of the real estate market supervision and early warning at home and abroad. This paper mainly introduces several typical real estate market supervision systems and the progress of domestic real estate early warning research, which lays the foundation for finding a reasonable early warning scheme. Secondly, this paper expounds the real estate cycle theory. The theory and elements of real estate early warning, the theory of cause of real estate market fluctuation, and the combination of economic principle and real estate industry lay a theoretical foundation for the following work, such as the selection of indicators. This paper finally selects the real estate market early warning means based on neural network. Firstly, the paper introduces the preparation work needed for early warning modeling, that is, data preprocessing, including the numerical definition of time difference analysis screening index and alarm degree, interval division; Secondly, the algorithm of BP neural network is selected, the principle of the algorithm and the steps of applying the algorithm modeling are introduced in detail. The combination of BP neural network and early warning system modeling and analysis is realized, and the principle of the empirical analysis is discussed. This article selected Tianjin real estate market as the sample to carry on the empirical analysis, first introduced the Tianjin city as the sample city and the division index time interval basis, described the Tianjin real estate industry development course; secondly, By querying Tianjin Statistical Yearbook and Tianjin Bureau of Statistics website, the detailed data of Tianjin real estate early warning index are obtained to ensure the authenticity and accuracy of the data. Finally, the data are analyzed and processed. The early warning index of Tianjin real estate market is screened, the model is modeled by BP neural network, the training and parameter determination of BP neural network are realized by MATLAB software, and the warning situation of Tianjin real estate market in 2012 is predicted. The conclusion that the market is running as "hot" is concluded. Finally, this paper summarizes the research results and conclusions of this paper, focuses on the analysis of the shortcomings of the early warning system, prospects for the future research direction, and puts forward some policy suggestions for the future development of early warning research from the perspective of the system.
【学位授予单位】：兰州交通大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP183;F299.23

【参考文献】