上海市办公楼需求量的神经网络组合预测

The Combined Forecast of the Demand of Office Building in Shanghai City Based on Neural Network

本文以上海市办公楼需求量为研究对象,将其看作城市常住人口、办公楼建设投资额、GDP规模、办公楼租金指数、人均可支配收入、办公楼空置率等六个参变量的非线性函数,并采用Logistic预测法、人工神经网络预测等方法对六个参变量进行预测,在此基础上,分别利用BP、RBF和ELMAN神经网络模型对上海市办公楼的需求量进行组合预测。预测结果对政府部门和房地产企业制定相关政策和决策有一定指导意义。

This paper studies the demand of office building in Shanghai. The demand is defined as the nonlinear function of six parameters, including permanent population of city, amount of investment on office building construction, the GDP, the rent index of office building, average available income for each person and vacancy rate of office buildings. Then, logistic forecasting and artificial neural network methods are used to calcu- late the value of those parameters. Based on the previous conditions and analysis, we forecasted the demand of office building in Shanghai by using BP, RBF, ELMAN neural network modes respectively, and concluded that the new demand, construction investment and rent price index would decrease slowly in 2008. Generally, government organization and real estate company can use the results for relative policy setting and decisionmaking