【作者】 汤红玲；

【导师】 黄映辉；

【作者基本信息】 大连海事大学 ， 企业管理学， 2005， 硕士

【摘要】 市场比较法、收益法和成本法是我国房地产估价的基本方法,它们在实际运用中存在着一个共同的问题:过于主观化,即依赖于房地产估价师的主观经验。这不仅使估价结果显得不可靠,更有可能引发道德风险,不利于整个房地产市场的规范与发展,因此,在房地产估价过程中建立定量分析的模型、减弱人为因素的影响,是解决这一问题的迫切需要。 市场比较法在其主要操作环节上没有确定统一的模式,很大程度上依赖于房地产估价师根据自己的经验做出判定,不同的房地产估价师对同一房地产进行评估可能会得出迥异的结论,且无法检验其准确性。将模糊综合评判方法用于可比实例的因素修正可减轻估价过程中的主观化问题。建立因素修正的FCE模型对可比实例的各因素进行修正,可将其交易价格修正到与估价对象因素相同或相近的价格。 收益法中过于主观化的表现主要在于资本化率的求取。现行资本化率的四种确定方法在实际操作中不能对影响资本化率的风险因素进行准确的界定。不同房地产估价师根据相同的情况得出的资本化率的细微差距会导致估价结果的较大波动。将人工神经网络用于确定资本化率,输入影响资本化率的主要因素,利用BP网络的自适应学习的特点进行反复训练和自我学习,可得出趋于稳定的输出,即资本化率。 过于主观化在成本法中的表现在于房地产客观成本和建筑物成新度的确定。本论文对现行确定客观成本的方法进行剖析,得出了客观成本应以实际成本为基础计算、而不应以同类房地产的社会平均成本计算的结论,并建立了求取客观成本的模型。同时将人工神经网络用于确定建筑物的成新度,以城乡建设保护部发布的《房屋完损等级评定标准》中所规定的影响建筑物成新度的12个因素作为输入,经过隐含层的反复训练和学习,得出趋于稳定的输出——成新度。 市场比较法、收益法和成本法的应用条件和使用范围是有差异的,如何根据不同的条件选用恰当的方法也是估价过程中必须关注的要点之一。更多还原

【Abstract】 Market Comparison Approach, Income Approach and Cost Approach are the three basic approaches in our nation’s Real Estate Appraisal. In practice, there exists a common problem in the approaches, that is, the appraisal processes are all too subjective, depending on the experience of the real estate appraisers too much. This not only will make the result of appraising unreliable, but also may result in the moral risk, and then go against on the development of the whole real estate market. So it’s in urgent necessary to set up models to make quantitative analysis in the appraisal process.In Market Comparison Approach, there isn’t a fixed model in the process. The appraisers make the result depending on, in a great degree, their own experience, so, upon the same real estate, different appraisers may make out different results. By using Fuzzy Comprehensive Evaluation (FCE) in modifying the factors of comparable property, that is, set a FCE model in the process of Market Comparison Approach can not only weaken the function of appraisers’ personal experience, but also put forward an objective result.In Income Approach, the process of getting the capitalization rate, a key factor in this approach, is too subjective, different appraisers may bring out different results according to their different experience. The small difference of capitalization rate may bring on the great difference of appraisal result. Using the Back-Propagation(BP) of Artificial Neural Net(ANN), putting in the main effect-factors of capitalization rate, and then the BP net will train thousands of times till the result go steady , that is the output of the net, capitalization rate.Cost approach puts up too subjectively by the confirmation of the objective-cost and the fresh degree of the subject property. Analyzing the method of confirming the objective-cost, it come to the conclusion that objective-cost should be calculated by the real-cost, but not the society average cost of the similar properties, and meanwhile, sets up a model for getting the objective-cost. The thesis uses Artificial Neural Net to confirm the fresh degree of the subject property, puts the 12 factors, which prescribed in ’Building Damage Grade Evaluation Standard’—a standard issued by the Construction Protection Part, that influence the building’s fresh degree into this ANN model. Byrepeated training and studying, we can get the steady output——fresh degree of thesubject property.There is some difference in the applying condition of Market Comparison Approach, Income Approach and Cost Approach. How to choose the right method by different condition is also one of the key points that real estates appraisers must notice in the appraisal process.更多还原