中文摘要：

以徐州为研究对象,运用灰色关联和BP神经网络预测住宅需求。结果表明：1）在岗职工平均工资、人均可支配收入、GDP、人均存款余额以及人均消费性支出等5项社会经济指标与住宅需求高度相关;2）BP神经网络的预测误差为6.18%,精度可以满足住宅需求预测的要求;3）徐州市住宅市场发展前景看好。

英文摘要：

Taking Xuzhou as the research object, the paper uses the gray correlation and BP neural network to forecast housing demand. The results show that： 1） Five socio-economic indicators are highly correlated with commodity housing demand, which include average wage of workers, per capita disposable income, GDP, per capita deposit balances and per capita consumption expenditure. 2） The prediction error of BP neural network is 6.18%, which can meet the requirement of housing demand forecast. 3） The housing demand of Xuzhou will show a good develpment prospect.