摘要
随着城市化进程的推进,公园资源成为许多楼盘选址的重要考虑因素,准确地衡量到公园距离对房价的影响对于房屋估价来说十分重要。特征价格模型难以非常有效地解决非线性多变量的情况,在衡量影响程度方面缺少说服力。利用神经网络解决特征价格模型在多变量非线性情况下拟合困难的问题,并与特征价格模型比较验证其拟合效果更优。采用根据平均影响值(MIV)算法原理改进的方法,保证在不产生异常值的情况下量化出到公园距离对房价的影响程度,结果表明到公园的距离越远则房价越低,并且公园绿地面积越大、绿化率越高对周围房价的影响越大。
With the advancement of urbanization,park resources have become an important factorin the location selection of many real estates.It is very important to accurately measure the impact of park distance on house prices for house valuation.The hedonic price modelcan not solve the nonlinear multivariable situation effectively,and it is not convincing in measuring the degree of influence.The neural network is used to solve the problem that thehedonic price model is difficult to fit in the case of multivariable nonlinearity,and itsfitting effect is better compared withthe hedonic price model.An improved method based on the principle of mean impact value(MIV)algorithm is used to ensure that the influence of distance to the park on house prices can be quantified without outliers.The results show that the longer the distance to the park,the lower the house price,and the larger the green area of the park and the higher the greening rate,the greater the impact on the surrounding house prices.
作者
李蔚民
牛卓颖
张英杰
LI Wei-min;NIU Zhuo-ying;ZHANG Ying-jie(School of Science,Beijing Forestry University,Beijing 100083;School of Economics and Management,Beijing Forestry University,Beijing 100083)
出处
《价格月刊》
北大核心
2020年第9期7-13,共7页
基金
国家自然科学基金青年项目“城市家庭对地方公共品的异质性偏好识别与应用研究”(编号:71603024)
中央高校基本科研业务费专项资金资助项目“中国的森林城市:实践经验、影响效果与量化评估”(编号:2019BLRD11)
北京林业大学“北京市大学生科学研究与创业行动计划”(编号:S201910022039)。
关键词
房价
特征价格模型
神经网络
公园
MIV算法
house price characteristics
price model
neural network
park
MIValgorithm
