指出了中国科技园区的经济发展呈现出高度的非线性特征,常规的经济预测手段很难对其作出准确的预测。支持向量回归实现了样本数据从数据空间到特征空间的映射,具备较强的非线性数据处理能力。提出了基于支持向量回归算法的中国科技园区经济发展预测思路,并建立了相应的数学模型。最后,对中国科技园区未来5年的发展状况作出了预测。
The paper first points out that the economic development of the state scientific and technological park region exists nonlinear characteristic to a great degree. The general economical forecasting methods become inefficient. Because it can carry out the sample data mapping from data space to feature space, support vector regression(SVR) has excellent ability to solve nonlinear problem. The forecasting model of economical development for the state scientific and technological park region based on SVR is proposed, For the purpose of testifying the validity of the model, the economic development of the state scientific and technological park region is forecasted in 5 years.