基于支持向量机的工程项目投资风险评价研究

Study on the Method of Projection Risk Evaluation Based on Support Vector Machine

提出一种工程项目投资风险评价的新方法—支持向量机评价法,由于支持向量机对样本数量的依赖性弱,通过学习有限的样本而建立的模型仍具有很强的泛化能力,具有比神经网络更好推广性能,因此在项目风险评价方面必然比神经网络具有更大的优势。实例研究也表明支持向量机方法更适合于投资项目风险评价的研究。

A novel projection risk evaluation method based on support vector machines （SVM） has been proposed in this paper. SVM has better generalization than BP neutral network with small sample data, hence naturally take more advantages in projection investment risk evaluation. The algorithm validation has been done with high technology projections in Fujian province and the experiment results support our conclusion.