【作者】 张园；

【导师】 武永祥；

【作者基本信息】 哈尔滨工业大学 ， 管理科学与工程， 2014， 博士

【摘要】 解决好城市居民的住房问题是一个庞大的系统工程，需要我们准确把握住房消费群体的需求特征，在此基础上制定更为有效和适宜的住房政策。近些年我国城市住房市场的供给与需求均十分旺盛，住房市场发展极为迅速，极大地促进了我国经济的繁荣和发展。与此同时，因对住房需求把握不清导致住房的建设和配置结构不尽合理，住房供求结构失调，出现大规模住房空置与供给不足现象并存，这是宏观数据和宏观研究难以解决的问题，需要对住房资源的微观配置进行分析。微观的住房资源配置是以住房商品与住房消费主体的匹配来体现，即居民个体或家庭与住房之间的匹配。基于上述背景，本文利用问卷调查获取的居民或家庭等个体微观数据开展居住联合选择研究，探索住房商品与住房消费者特征的匹配关系，更加真实的反映和解释现实住房需求特征，帮助政府和住房供给者更为准确的把握居民的住房需求特征，进而优化住房供给结构，提高住房资源的配置效率。本文对住房联合选择的相关研究基础进行分析，包括对住房需求特征、居住联合选择等概念进行界定，对居住联合选择相关理论基础进行分析，在此基础上结合学术研究历史和现实居住选择行为特征对居住联合选择进行理论分析，构建包含首次购房时间与住房权属联合选择、住房权属与建筑特征联合选择以及住房邻里与区位特征联合选择内容的理论模型，并对上述内容的研究因素进行了分析，此为基础确定需要收集的基础数据，设计相应的调查问卷，选取北京、上海、哈尔滨和成都作为样本收集城市，进行问卷发放和回收，总计回收微观样本1581份。针对首次购房时间与住房权属的联合选择问题，本文基于生存分析方法构建Life Tables和Cox模型，对潜在首次购房者首次购房生存时间的分布函数和首次购房事件发生概率的影响因素进行了纵贯分析，利用问卷调查获取的1062份回溯性微观数据和官方公布的宏观数据对理论模型进行标定，得到了潜在首次购房者首次购房生存时间的分布函数，分别得到了个体特征、宏观经济和房地产市场环境、货币政策、二手房交易税收政策、差别化的信贷政策、土地政策以及生命历程重要事件等因素对首次购房事件发生概率的影响。针对住房权属与建筑特征类别的联合选择问题，本文在对贝叶斯网络基本知识介绍的基础上，利用北京、上海、成都、哈尔滨四个城市的1136份样本数据，从个体人口-社会经济属性和住房特征心理偏好程度选取研究变量，应用K2算法进行贝叶斯网络结构学习，应用最大似然法进行网络参数学习，构建了住房权属与建筑特征类别联合选择的贝叶斯网络模型。针对所构建的模型并借助于联合树推理引擎进行贝叶斯网络推理，得到住房权属与建筑特征类别类别联合选择的人群特征诊断推理结果。针对居住区位与邻里特征的联合选择问题，本文构建包含居住区位特征（包括自然环境、交通、教育、商业服务资源）和住房邻里特征（包括小区绿化率、房屋均价、物业管理水平、建成年代）变量的指标体系，对哈尔滨141个样本小区进行层次聚类，得到包含特征差异显著的五类联合居住区位选择集。从个体人口-社会经济属性、居住需求、工作通勤需求和居住区位特征心理偏好程度四方面选取变量作为居住区位选择的驱动因素，对315个个体样本进行探索性分析和MNL模型分析，得到了各变量对居住区位选择的驱动效果，验证了宜居性特征的陈述性心理偏好与实际选择的显示性偏好的一致性关系。本文对基于三条主线构建的三个居住联合选择模型结果进行分析、总结，并对住房需求特征不同类别的关联关系进行基于最优尺度变换的多重对应分析，验证理论模型的合理性，并揭示我国城市居民的住房需求微观特征，住房与选择主体的匹配关系，此基础上分别对政府部门、房地产开发企业和住房消费者提出相关对策建议。更多还原

【Abstract】 Solving the housing problem is a complicated systematic engineering. To makerelevant housing policies requires us to understand the characteristics of the housingconsuming group. In recent years, both the supply and the demand in China’shousing market have shown their prosperity. China’s economy was boosted by therapid development of housing market. Meanwhile, the insufficient study on housingdemand has resulted in irrational construction and supply-demand structureimbalance. Many houses are left vacant while others are short for demand. This canhardly be solved by macroscopical study but microcosmic study on housing resourceallocation. Microcosmic housing resource allocation is reflected by the match ofhousing commodity and housing consumer, viz. the match of individual resident andhouse. Based on the above mentioned conditions, this dissertation studies themicro-data obtained from questionnaires, explores the relationships betweenhousing commodity and housing consumer’s characteristics, reflects and explainshousing consuming characteristics, helps the government and housing supplierscatch the residents’ housing consuming mentality, and improves the efficiency ofhousing resource allocation.The dissertation analyses the study basis related to housing choice, includingthe definition of housing, housing demand, housing choice, and the theory basis ofhousing choice study. Then, based on the study structure, analyses and selects thevariables of the joint choice of housing tenure and building characters, residentiallocation choice and the time of first homeownership, decides the data to be collected,and designs related questionnaire. Beijing, Shanghai, Harbin and Chengdu arechosen to be the target cities, and finally return1581micro-samples.For the joint choice of first-time homeownership and housing tenureproblemthis dissertation used Life Tables and Cox model of Survival AnalysisMethod, longitudinally analyzed the distribution function of potential first-timehousing consumers’first-time homeownership survival timeand factors affecting theoccurrence probability of first-time homeownership. It introduced the SurvivalAnalysis Method, combined research contentsto established theoretical model andthen used the questionnaire survey to obtain1062retrospective micro data andofficial released macro data of theoretical model calibration, hence acquiredthedistribution function of the potential first-time house consumers’ first-timehomeownership survival time. In addition, it acquired factors affecting theprobability of occurrence of the first event of the purchase, such as individualcharacteristics, macro-economy and real estate market environment, monetary policy, transaction tax policyofsecond-hand housing, differentiated credit loanpolicy, land policy, and important events in human life.For the joint choice of housing tenure and building characteristics’ categoryproblem this dissertation chooses variables from the aspect of individual demograticand socio-economic attribute and residential preference among1136samples fromBeijing, Shanghai, Chengdu and Harbin, and builds up a Bayesian Networkmodelbased on the knowledge of Bayesian Network. The model is built up bystructure study using K2algorithmand parameters studu using Maximum LikelihoodEstimation (MLE).The result of consumer groups characteristics diagnose inferenceof housing categories joint choice are finally obtained by using and Joint TreeInference Engine.For the joint choice of residential local and neighborhood characteristicproblem this dissertation constructs an index system from the aspects ofenvironment, education, business, communications and neighborhood andcommunity.141community samples in Harbin city are classified by the above fivecategories and make up a choice set. Variables such as individual demogratic andsocio-economic attribute, housing demand, working commuting demand andlocation preference of amenity characters are selected as the driving factors of jointlocation choice. The driving effect on location choice of each variable is obtainedafter the exploratory analysis and MNL model analysis. The consistency of statedresidential preference of amenity characters and revealed preference of timingchoice is also testified.In this dissertation, three joint selection model results aresummerized and therelationship between housing demand characteristics associated with differentcategoriesis analyzedwith multiple correspondence analysis based on the optimalscale transformation, verifying the rationality of the theoretical model and revealingthe microscopic features of urban residents’ housing demand and matchrelationships with those who choose. Finally, this dissertationputs forward relevantsuggestions to the government departments, sto and housing consumers.更多还原