城市商业地产是城市商贸水平的空间集中体现，是城市商业发展繁荣程度的反映， 也是城市化发展水平的表现。近年来电子商务以其诸多优势吸引大批消费者，替代部分 实体市场，导致城市商业地产实体商铺遭受巨大冲击。在线下、线上融合的新消费环境 下，研究城市商业地产面积规模对城市土地资源的科学规划和商业资本的合理利用都有 重要意义。另外新型冠状病毒造成的疫情使得国内城市商业活动短期内基本停滞，造成 大量商铺关门歇业。合理的城市商业地产面积规模是城市规划部门与企业高层决策者急 需破解的重要课题。 本文通过梳理城市商业地产预测方法和商业地产文献，提出本文研究的商业地产研 究范围，包括零售、餐饮、娱乐等用于商业的房地产，不包括办公楼。以城市商业地产 面积规模为研究对象，分别以西部重要省会城市西安市和成都市为例进行实证研究，重 点探究电子商务因素对西部地区城市商业地产面积规模的影响。根据 2000 年至 2018 年 商业地产数据，将相关性分析与灰色关联度方法相结合，确定了城镇人口、恩格尔系数、 地区生产总值、第三产业产值、商业营业用房投资额、社会消费品零售总额、城乡居民 储蓄存款、城镇居民人均可支配收入以及电子商务交易额和互联网用户规模 10 个与城 市商业地产相关性较高的指标体系，进而利用多指标 BP 神经网络，构建了新增城市商 业地产面积规模预测模型。 基于西安与成都商业地产发展现状，2000 年至 2018 年，西安市新增商业地产面积 整体呈上升趋势，增幅较快，年增长率为 26%；成都市呈波动增长趋势，增速相对较慢， 年增长率为 14%。成都市年新增出售商业地产面积均高于西安市，每年新增出售规模差 距在 28—140 万平方米之间。根据 BP 神经网络模型预测，发现电子商务因素对城市商 业地产面积规模影响较大。2019 年至 2025 年，引入电子商务因素前后，西安年新增出 售商业地产面积规模都呈平缓增长趋势，且每年新增出售面积差距保持在 10—31 万平 方米之内；未引入电商因素时商业地产面积规模和增速都大于引入后的面积规模，其中， 未引入电商因素的年增长率为 3.65%，引入后年增长率为 2.75%；成都年新增出售商业 地产面积规模呈倒“U”发展趋势，引入电商因素前后每年新增面积规模差距在 15—60 万平方米之间；未引入电商因素预测的面积大于引入后的面积规模，但增速明显小于引 入后，其中，未引入电商因素的年增长率为 4.94%，引入后年增长率为 7.06%。预测结 果表明，2019 年至 2025 年，成都市每年新增出售商业地产面积规模和整体增速都大于ii 西安市。未引入电商因素时成都每年新增出售面积比西安市高 64—200 万平方米，引入 后成都每年新增出售面积比西安高 113—200 万平方米。 通过对比我国西部两个省会城市西安市和成都市商业地产面积规模在未来短期的 需求量发现，未来西部城市商业地产面积规模还将持续增加，达到一定规模后在 2025 年开始出现下降趋势，但规模总体少于未引入电子商务影响下的需求量，网购消费的介 入使得城市商业地产面积规模明显减少。 关键词：商业地产，面积规模，电子商务，BP 神经网络，预测iii Abstract Urban commercial real estate is the concentrated embodiment of urban commercial level of space, reflects the prosperity of urban commercial development, but also the performance of urbanization development level.In recent years, e-commerce attracts a large number of consumers with its many advantages and replaces the market of some physical commodities. Electronic commerce in recent years, with its many advantages to attract a large number of consumers, and replace the part of the physical commodity markets, particularly in the retail industry occupies the important position, lead to hit city commercial real estate entities shops, and with the emergence of "new retail sales model", online fusion integration of modern logistics development will be more adaptable to market needs.In the new consumption environment, it is of great significance to study the development of commercial real estate area scale required by the city in the future for the planning of urban land resources and the utilization of social capital. In addition, the outbreak caused by novel coronavirus pneumonia caused the domestic urban commercial activities to basically stop in a short period of time, and the reasonable size of the urban commercial real estate area has become an important issue urgently needed to be solved by the urban planning department and the high-level decision makers of enterprises. By combing and summarizing the prediction methods and commercial real estate literature of urban commercial real estate, this paper proposes the research scope of commercial real estate involved in this paper, that is, the narrow commercial real estate, including retail, wholesale, catering, entertainment and other commercial services, excluding office buildings.Taking the area scale of urban commercial real estate as the research object, Xi 'an and Chengdu, two important provincial capitals in western China, were taken as examples to conduct empirical research, focusing on the influence of e-commerce and other factors on the area scale of urban commercial real estate in western China.According to the data from 2000 to 2018, commercial real estate, combining correlation analysis and grey relational analysis, determines the urban population, GDP, tertiary industry output value, engel coefficient, commercial buildings for business investment, total retail sales of social consumer goods, urban and rural residents savings, per capita disposable income of urbaniv resident, and e-commerce transactions and the number of Internet users 10 high correlation with the city commercial real estate index system, then using multi-index BP neural network, build the size of the urban area of commercial real estate short-term prediction model in the future.The annual increase of commercial real estate sold in Chengdu shows an inverted "U" development trend. The annual increase of commercial real estate size before and after the introduction of e-commerce is between 150,000 and 600,000 square meters. The predicted area without e-commerce factor is larger than the area scale after the introduction, but the growth rate is obviously smaller than that after the introduction Based on the development status of commercial real estate in Xi 'an and Chengdu, from 2000 to 2018, the newly added commercial real estate area in Xi 'an has been on the rise with an annual growth rate of 26%. Chengdu shows a trend of fluctuating growth, which is relatively slow, with an annual growth rate of 14%.The area of commercial real estate newly sold in Chengdu is higher than that in Xi 'an, and the gap of newly sold scale is between 280,000 and 1.4 million square meters.Through calculation and simulation, it is found that e-commerce factors have a great influence on the area scale of urban commercial real estate.From 2019 to 2025, before and after the introduction of e-commerce, the annual increase in the area of commercial real estate sold in Xi 'an shows a flat growth trend, and the gap of the annual increase in the area sold remains within 100,000-310,000 square meters. When the factor of e-commerce is not introduced, the area scale and growth rate of commercial real estate are both larger than the area scale introduced.Among them, the annual growth rate without the introduction of e-commerce is 3.65%, and the annual growth rate after the introduction of e-commerce is 2.75%.Among them, the annual growth rate without the introduction of e-commerce is 4.94%, and the annual growth rate after the introduction of e-commerce is 7.06%.The forecast results show that from 2019 to 2025, the area size and the overall growth rate of new commercial real estate sales in Chengdu will be higher than that in Xi 'an.When the e-commerce factor is not introduced, the annual newly sold area in Chengdu is 640,000 to 2 million square meters higher than that in Xi 'an, and the annual newly sold area in Chengdu is 1.13 million to 2 million square meters higher than that in Xi 'an. By comparing the two western provincial capital city of xi 'an and Chengdu area the sizev of the commercial real estate short-term demand found in the future, the future scale of commercial real estate area in western cities will continue to increase, after reaching a certain scale, in 2025 started to decline, but the overall scale is less than the demand under the influence of the introduction of e-commerce, online consumer involvement makes city commercial real estate area scale decreased significantly. Key words:Commercial real estate，Area scale，Electronic Commerce，BP neural network，Forecastvi 目 录 第一章 绪论...1 1.1 研究背景.............................1 1.2 研究目的与意义.................2 1.2.1 研究目的.............. 2 1.2.2 研究意义.............. 2 1.3 研究思路与内容.................3 1.3.1 研究思路.............. 3 1.3.2 研究内容.............. 6 1.4 研究创新点.........................6 第二章 相关理论与文献综述......8 2.1 城市商业地产定义.............8 2.2 预测方法.............................8 2.3 文献综述...........................12 2.3.1 国外研究............ 12 2.3.2 国内研究............ 13 2.4 本章小结...........................15 第三章 城市商业地产面积规模预测模型构建.......16 3.1 城市商业地产面积规模影响因素指标体系确定.........................16 3.1.1 影响因素概述.... 16 3.1.2 影响因素指标筛选方法................... 19 3.1.3 西安市商业地产面积规模影响因素指标确定.............. 20 3.1.4 成都市商业地产面积规模影响因素指标确定.............. 27 3.2 基于 BP 神经网络的城市商业地产面积规模预测模型...............33 3.2.1 BP 神经网络预测模型原理.............. 33 3.2.2 BP 神经网络预测模型构建.............. 35 3.2.3 实证研究城市的选择....................... 37 第四章 西安市商业地产面积规模预测...................38 4.1 影响因素指标处理...........38 4.2 未引入电子商务因素的西安市商业地产面积规模预测模型构建............................41 4.2.1 网络训练阶段.... 41vii 4.2.2 网络测试仿真阶段........................... 41 4.2.3 预测阶段............ 42 4.3 引入电子商务因素的西安市商业地产面积规模预测模型构建.43 4.4 西安市商业地产面积规模预测结果分析......44 第五章 成都市商业地产面积规模预测...................46 5.1 影响因素指标处理...........46 5.2 未引入电子商务因素的成都市商业地产面积规模预测模型构建............................49 5.3 引入电子商务因素的成都市商业地产面积规模预测模型构建.51 5.4 成都市商业地产面积规模预测结果分析......52 第六章 西安与成都商业地产面积规模预测比较研究..........................54 6.1 商业地产发展现状对比...54 6.1.1 商业地产投资方面........................... 54 6.1.2 商业地产空置面积........................... 55 6.2 预测结果对比...................56 6.3 对比结果原因分析...........58 6.3.1 经济发展，投资渐趋平稳............... 58 6.3.2 人口增加，消费不断扩大............... 59 6.3.3 电商竞争，业态调整转变............... 60 第七章 结论与展望....................62 7.1 研究结论...........................62 7.2 研究建议...........................63 7.3 研究局限性.......................63 7.4 研究展望...........................64。。。。。。以下内容略