城市建设用地规模预测模型探讨 ——以河北省邢台市为例

国内众多专家学者对确定城市合理用地规模的方法上进行了多角度探索,本研究有益借鉴这些成果,采用多元线性回归分析和BP神经网络2种较成熟的预测模型,选取城市建设用地面积、地区生产总值、固定资产投资、城市人口数、年末道路面积、城市园林绿地面积、商品房平均售价为观测指标,分析预测河北省邢台市的建设用地规模。结果表明,预测结果并不理想,预测值与实际值的误差始终存在,并随时间推移不断扩大,深入分析预测结果产生偏差的原因,即忽视了非理性因素对城市建设用地规模的影响。建议引入行为金融学理论,将研究对象具化到个人,研究土地市场管理者和投资者在决策过程中的认知、行为偏差,分析城市建设用地面积持续蔓延的非理性动力,应用行为组合理论和资产定价模型对传统研究范式进行有益补充和修正,进而更科学地把握城市用地空间形成机制及演变规律。

Study on Prediction Model of Urban Construction Land Scale A Case Study of Xingtai in Hebei Province

urban land by using reasonable methods. In order to analyze and forecast the scale of construction land of Xingtai City in Hebei Province, the author chose 2 mature prediction models： the multiple linear regression analysis and BP neural network, which were very effective tools in the similar study over the world. But the prediction results of 2 models were not satisfied, a depth analysis of the reasons showed that it was the ignoring of the irrational factors on the impact of urban construction land. So the author advised to use the behavioral finance theory to study the land market participants＇ cognitive biases in decision process, improving the validity of the prediction model, which could be more scientific to urban space formation mechanism and evolution.