基于生长曲线、ARIMA模型和神经网络的企业数短期预测和比较——以深圳为例

Short-Term Forecasts of Enterprises' Number Based on the Growth Curve,ARIMA Models and Neural Networks with the Comparison on the Three Methods:A Case Study in Shenzhen

企业是市场的主体和产业的载体,目前阶段企业数的增长是衡量地区创新活力的重要标志。基于1988～2008年深圳企业数据,运用Logistic生长曲线、ARIMA模型和动态神经网络对2009～2011年深圳企业数进行短期预测,论证了企业数短期预测的可行性,并在预测结果的基础上分析了三种预测方法的优缺点。企业数短期预测的实现可以为相关政策的制定提供参考。

At current stage, the growth of enterprises＇ number is an important symbol to measure regional inno- vative vigor, since enterprises are the main body of market and carriers of industry. By using the logistic growth curve, ARIMA model and dynamic neural network, the short - term forecast of Shenzhen enterprises＇ number in 2009 -2011 could be made based on the data in 1988 -2008. And of which, the feasibility is demonstrated. The advantages and disadvantages of three prediction methods are analyzed on the basis of predicted results. The reali- zation of the short - term forecast of enterprises＇ number could provide references for policymaking.