青岛市房地产市场预警系统建模及其实证研究
本文关键词:青岛市房地产市场预警系统建模及其实证研究,由笔耕文化传播整理发布。
中国海洋大学
硕士学位论文
青岛市房地产市场预警系统建模及其实证研究
姓名:周琦
申请学位级别:硕士
专业:管理科学与工程
指导教师:张勤生
20080530
青岛市房地产市场预警系统建模及其实tlF研究
青岛市房地产市场预警系统建模及其实证研究
摘要
房地产业在整个国民经济体系中属于先导性、基础性产业,处于主导产业地位,其存在着明显的周期波动规律。起伏过大的波动与房地产经济的持续健康稳定发展相矛盾,但是目前,我国房地产市场运行机制不甚完善,还没有形成合理、有序、竞争、高效的市场运行体系,房地产市场存在信息传递不畅、信息数据失真、市场行情展示手段落后和市场交易网络封闭等?系列问题。因此,研究房地产预警系统,设置房地产预警指标体系,系统、科学、准确地确定房地产安全区域,成为有关决策部门亟需解决的重大现实问题,又是学术界需要深入研究的重大理论问题。
针对这种情况,本文提出了房地产市场预警系统模型研究,为促进房地产业的健康发展提供一定的理论依据。并在此基础上,对青岛市房地产市场进行了实证研究。
本文通过研究,得到的研究成果及研究结论主要有以下几个方面:
(1)基于对国内外房地产预警研究现状的分析,归纳总结了房地产预警的基本概念、基本原则和基本方法。在对国内原有各类房地产指标体系进行研究的基础上,确定房地产预警指标。
(2)房地产作为社会经济系统的一个子系统,具有非线性复杂系统的特性。本文立足于解决房地产系统的非线性问题,建立更为先进科学的房地产预警系统,避免房地产市场的非常态波动,促进房地产市场的持续、健康、稳定发展。在现有研究的基础上,系统地分析了房地产预警的特点及功能特征,对房地产预警过程中的关键预警指标进行了辨识、预测、诊断、监测和控制,构建了具有理论性和实践性的房地产预警系统,为解决房地产预警问题提供了依据。
(3)本文介绍的房地产预警系统,以神经网络理论和房地产预警理论为基础,构建了预警模型。利用神经网络在预测和模式识别领域的成熟运用,重点探
青岛市房地产市场顶警系统建模及其实证研究
讨基于神经网络理论的房地产预警的模型和方法,并利用该模型形成了房地产市场预警体系。
(4)在建立了房地产预警指标体系的基础上,本文提出了基于LVQ—RBF神经网络的房地产预警模型,该模型克服了传统预警方法的不足,具有高度的并行性和全局性,提高了房地产预警系统的非线性、自学习性、自适应性及大规模并行分布知识处理的能力,具有较高的精确度和适用性。
(5)本文依据前期研究理论成果,对青岛市房地产风险预警进行实证分析,依据技术可能、经济合理、操作可行等原则,最终形成综合预警分析结论。预警分析的结果与青岛房地产发展的实际情况基本吻合,表明本项研究所建立的房地产预警模型系统有效可行,理论分析充分,实用价值高,为指导和调控房地产市场提供了科学依据。
关键词:房地产;神经网络:系统建模;学习矢量化;预警;预测【I
metsyboadgn’Q
青岛市房地产市场预警系统建模及je实证研究
ModeIingforRealEstateForecastingandEarlYWarning
Systemandiitssl:mpEmlilrlicaIIReseaesearrchhIinUingdao
Abstraot
Realestateindustry,whichisplayingaleadingroleinthedevelopmentofnationaleconomy,showsmoreandmoreobviouslythecharacteristicofcyclefluctuation.Itsexcessivefluctuationcontradictswith也esustainable,healthy.andstabledevelopmentoftherealestateeconomy.Currently,thereisstillmuchworktodotoperfecttherealestatemarketoperationmechanismofChina,tOformrational,orderly,competitiveandefficientmarketoperationsystemandtosolvethecurrentproblemsoftheunsmoothtransferringofinformation,thedistortionofinformationdata,thebackwarddisplayingmeansofmarketquotationandtheblockingofthemarkettransactionnetwork.IthasbecomeahotpointintheacademiccyclesandarOI/Sesintensiveconcernsoftherelatedpolicy?makingapartmentstodoresearchonrealestateearlywarningsystem,tosetuptherealestateearlywarningindexsystemandsystematically,scientificallyandaccuratelydefinethesecureregionofrealestate.
Toaddressthisproblem,theresearchofearlywarningmodelsofrealestatemarketisputforwardinthisdissertationwhichwillmakeitstheoreticalcontributiontothepromotionofthehealthydevelopmentofrealestateindustry.Onthebaseofthisresearch,empiricalstudyisalsodoneontherealestatemarketofQingdao.
Basedontheresearchabove,themainfindingsandtheconclusionsarementionedasfollows:
Firstofall,basedontheanalysisofthecurrentresearchesathomeandabroad,thisdissertationsummarizesthebasicconcepts,principlesandmethodsofforecastingandearlywarningofrealestateeconomy.Onthebaseoftheanalysisofthedomesticrealestateindexsystems,thisdissertationselectsandfixesonitsindexesofearlywarning.
Secondly,asasubsystemofthesocialandeconomicsystem,realestateshowscomplexnon—linearcharacteristics.Thisdissertationaimsatsolvingthenonlinearproblemoftherealestatesystemandestablishingamoreadvancedandscientificearlywarningsysteminordertopreventtherealestatemarketfromtheabnormalfluctuationandtomaintainthesustainable,healthyandstabledevelopmentofthereallII
estatemarket.On山efoundationofthepresentresearch,thedissertationsystematicallyanalysesthecharacteristicsandfunctionsoftherealestateearlywarning,makeanidentification,prediction,diagnosis,monitoringandcontrolofthekeyindexintheprocessofrealestateearlywarningandconstructatheoreticallyandpracticallyfeasiblerealestateearlywarningsystemwhichprovidesthebasisforthesolutionoftherealestateearlywarningproblems.
Thirdly.basedonneuralnetworkstheoriesandrealestateearningwarningtheories,thisdissertationintroducesitsforecastingandearlywarningsystemanddevelopsitsmodelofforecastingandearlywarning.ByutilizingtheneuralnetworkswhichiSmaturelyappliedinthefieldofforecastandmodelrecognition,thedissertationputsitsemphasesontheresearchofthemodelsandmethodsofrealestateforecastingandearlywarning.Andbasedonthesemodelsthisdissertationdevelopsitsownforecastingandearlywarningsystemofrealestatemarket.
Fourthly,basedontherealestateearlywarningindexsystem,thisdissertationdevelopstheLVQ..RBFneuralnetworksmodelofforecastingandearlywarning.Withhighparallelism,globalsuperiority,accuracyandapplicabilitythismodelhasovercomethedeficiencyoftraditionalearlywarningmethodsandhighlyimprovestherealestateearlywarningsystem’Snon-linearity,selfstudyingability,selfadaptabilityandtheabilitytoprocesslarge—scaleconcurrentlydistributedknowledge.
Finally,basedontheprevioustheoreticalresearchfindings,thisdissertationempiricallyanalysesoftherealestatemarketinQingdaoandformsitscomprehensiveearlywarningtheoriesandmethods.SincetheconclusionoftheanalysisofearlywarninginthisdissertationisinaccordancewiththepracticaldevelopmentoftherealestateinQingdao,therealestateearlywarningsystemestablishedinthisresearchisprovedtobefeasible,withfulltheoryanalysisandgoodpracticalvalue,andprovidescientificfoundationforguidingandcontrollingtherealestatemarkets.
KeyWords:RealEstate,,NeuralNetworks,SystemModeling,LearningVectorQuantization,EarlyWarning,Forecasting
本文关键词:青岛市房地产市场预警系统建模及其实证研究,由笔耕文化传播整理发布。
本文编号:132522
本文链接:https://www.wllwen.com/jingjilunwen/fangdichanjingjilunwen/132522.html