【作者】 申文金；

【导师】 夏建国；

【作者基本信息】 四川农业大学 ， 土地资源管理， 2007， 硕士

【副题名】以德阳市为例

【摘要】 随着我国城市化进程的推进，城市自身的近域推进和广域扩展速度加快，城市边缘区(urban fringe)成为变化最大、最迅速、最敏感的地区，城市边缘区的土地开发利用和土地交易活动异常活跃。城市边缘区是城市与农村相互结合的特殊经济地理单元，是城市快速扩张的重点区域，受城市与农村的双重影响，使得城市边缘区的土地价格低于城区的土地价格，同时，又高于乡村的土地价格。城市边缘区是土地价格研究中的薄弱点，土地市场的快速、健康发展迫切要求对该区进行地价评估，开展中小城市边缘区土地价格评估研究对城市边缘区土地管理、土地交易、土地价格管理等具有一定的理论意义和现实意义。本文界定研究范围为中小城市边缘区土地，包括内缘区土地和外缘区土地，以四川省德阳市为例，对中小城市边缘区的土地价格评估方法进行了研究。在内缘区土地价格评估中，对内缘区土地价格的影响因素进行了分析，在此基础上，提出了中小城市内缘区土地价格评估指标体系，考虑到土地价格与影响因素之间的非线性关系，为了克服以往传统方法评估的局限性，引入了BP神经网络模型，对内缘区土地价格于影响因素之间高度的非线性关系进行模拟，以便科学的评估内缘区土地价格；针对中小城市边缘区中外缘区土地的特性，对外缘区土地的价格构成进行了分析，提出了各构成部分的价格内涵，在分析价格内涵的基础上，选用了科学的方法对外缘区土地的质量价格、社会保障价格、社会稳定价格和景观生态价格进行了评估。主要有以下几个方面：(1)针对内缘区土地价格的特性建立了一套内缘区土地价格评估指标体系，在上述研究分析的基础上，对BP神经网络应用于内缘区土地评估的可行性进行分析，进而引入BP神经网络并构建内缘区土地价格评估模型。(2)对德阳市内缘区土地价格进行调研和分析。运用建立的评估模型，将调研来的128个样本数据经过筛选后分成训练样本和测试样本两组，30个训练样本和6个测试样本，进而验证模型的有效性，获得了理想的模拟结果，6个评估模型测试结果与原始样本真实值进行对比，最小的相对误差为0.43％，最大的相对误差为0.84％，平均相对误差为0.58；最小模拟精度为99.16％，最大的模拟精度为99.61％，平均模拟精度为99.42％，结果表明该模型有较强的实用价值。(3)对德阳市外缘区土地价格进行调研和分析。运用土地价值理论，对外缘区土地价值构成进行分析——完整的外缘区土地价格由土地质量价格、社会保障价格、社会稳定价格和景观生态价格构成。根据外缘区土地价格的四个构成部分，建立了德阳市的外缘区土地价格评估体系。并对其各价格构成部分进行测算，德阳市外缘区的土地质量价格为16.50元／m~2；社会保障价格为140.35元／m~2：社会稳定价格为12.61元／m~2；景观生态价格为10元／m~2，外缘区土地价格等于上述4种价格之和即179.46元／m~2。本文运用BP神经网络对内缘区土地的价格评估进行研究，但是由于内缘区土地评估本身涉及的因素众多复杂，包括样本数据的采集、一些不能避免的不可定量化因素以及不可预测因素的存在等，BP神经网络内缘区土地价格评估模型还有待进一步提高与完善。另一方面对外缘区土地价格各构成部分的评估方法中一些数据及参数的选取需要进一步科学化。更多还原

【Abstract】 During the course of urbanization and industrialization rapidly pushing on, urban fringe is becoming more and more peculiar economic and geographic unit. In urban fringe of middle and small city, the land transaction is frequently, the land price type is diversified. And the change of land uses is acute in urban fringe. Affected by development of city and countryside, city fringe area is a peculiar economic and geographical unit, and land price of urban fringe is higher than urban land, and lower than rural land. However, researches on land price of the urban fringe area are not enough, though the healthy development of the land market requires evaluating the land price within this area.In this paper, we mainly researched on the appraisal method of urban fringe land price in middle and small city include inner land and outer land, in the case of De Yang. The paper studies and analyses many factors with respect to inner land price according to actual situation in land market, and established inner land appraisal guide line. Based on the analysis of De Yang’s urban planning in the next 10 years, the paper shows the Characteristic Index for special tracts and uses the improved BP Neural Network to train and simulate the realistic price in inner land.On the other hand, the paper analyzed the structure of outer land price in urban fringe and evaluates each constitutions of outer land price, based on the theoretical basis of land price. This paper mainly brings three ideas:Emphasis the effect of urban planning and shows the Characteristic Index which affect the price of inner land. Based on the analysis of De Yang’s urban planning in the nest 10 years, the article shows the Characteristic Index for special tracts and uses the improved BP Neural Network to train and simulate the realistic price in inner land. (1) The paper studies many factors with respect to inner land price according to actual situation in land market, and established inner land appraisal guide line. On other hand, this paper analyses the feasibility of BP neural networks which used to appraising the land price of inner land, and expounds Genetic Algorithms based on neural networks and founds the inner land appraisal model.(2) The thesis divide a large number of data acquired by investigation and research into two groups——training sample and testing sample to verify the model, and applies the appraisal model founded. And then the paper analysis the differences between the test results of the appraisal model and the actual values, the minimal relative error is 1.22%, and the maximal relative error is 3.58%. The paper finally gets ideal results, which indicates the model has good scientific and practical value.(3) The article analysis the structure of land price of outer land in urban fringe according to theoretical basis of land price, investigation in outer land of urban fringe in De Yang. The comprehensive outer land price evaluation should also include the price of land quality, social security, social stability and landscape ecology. With the suitable method, this article gives the value of land quality, social security, social stability and landscape ecology. The value of land quality is 16.50 yuan/m~2, the value of social security is 140.35 yuan/m~2, the value of social stability is 12.61 yuan/m~2, and the value of landscape ecology is 10.00 yuan/m~2. In the paper we construct an outer land price evaluation system according to the four different values.The paper uses the BP neural networks to evaluate the land price of inner land, but the impact factors which affect land appraising are various and complex, including data acquisition of sample; the un-quantitative factors which are inevitable and theunpredictable factors etc. So the appraisal model based on BP neural network need improvement and perfect. On the other hand, the parameters in each appraising methods for outer land should be more scientific.更多还原