企业自主创新项目风险评价模型与实证研究

摘要

For solving the high dimensional and nonlinear problems of enterprise independent innovation project risk evaluation, an intelligent model based on principal component analysis ( PCA), genetic algorithm (GA) and artificial neural network ( ANN), i.e. PCA - GA - ANN, is proposed in this paper. In this hybrid approach, PCA is used for feature extraction in order to reduce the model complexity of ANN and improve the speed of ANN, GA is used to train the weights of ANN to constitute a GAANN model, and then the extracted principal components are introduced into GAANN to obtain the evaluation results. The empirical results reveal that PCA - GA - ANN model has understanding forecasting ability. Compared with the standard back- propagation artificial neural network (BPANN), PCA- GA- ANN has some superiority in predicting accuracy.%针对企业自主创新项目风险评价中的高维、非线性问题,提出了一种基于主成分分析和遗传神经网络的企业自主创新项目风险评价方法.该方法利用主成分分析对企业自主创新项目风险评价体系进行特征提取,利用遗传算法直接训练神经网络的权重形成遗传神经网络,特征提取后的综合主成分指标进入遗传神经网络的智能评价系统.实证结果表明,该方法具有较好的泛化能力,与标准BP神经网络方法相比,该方法具有明显的优势.