| 神经网络在高技术项目投资风险评价中的应用 |
| |
| 引用本文： | 张春梅. 神经网络在高技术项目投资风险评价中的应用[J]. 河南科技大学学报(自然科学版), 2004, 25(4): 36-38 |
| |
| 作者姓名： | 张春梅 |
| |
| 作者单位： | 安庆师范学院,信息科学与工程学院,安徽,安庆,246011 |
| |
| 基金项目： | 安庆师范学院科研基金资助项目 (2 0 0 3yiy0 7) |
| |
| 摘 要： | 在对高技术项目投资风险因素分析的基础上，建立了能够预测项目投资风险的三层Levenberg-Marquardt Backpropagation Neural Network(LM-BP)人工神经网络模型。采用此人工神经网络模型，隐含层和输出层传输函数皆为purelin,各层内的所有权重相等，偏置也相等，模型避免了“过拟合”现象的发生。预测结果表明，该神经网络模型稳定可靠，所获得的结果是令人满意的。
|
| 关 键 词： | 神经网络 高技术项目 投资 风险 |
| 文章编号： | 1672-6871(2004)04-0036-03 |
| 修稿时间： | 2004-04-30 |
Application of Neural Network to Investment Risk Evaluation of High-technical Projects |
| |
| ZHANG Chun-Mei. Application of Neural Network to Investment Risk Evaluation of High-technical Projects[J]. Journal of Henan University of Science & Technology:Natural Science, 2004, 25(4): 36-38 |
| |
| Authors: | ZHANG Chun-Mei |
| |
| Abstract: | Based on the analysis of the investing risk factor on high-technical projects, it has been found that three layers neural network model with purelin transfer function by Levenberg-Marquardt Backpropagation Neural Network (LM-BP) can be used to predict investment risk of the projects. In neural network model, the transfer functions of hide layer and output layer are both purelin function. In all of layers, each weight is equal and so is each bias. A problem of overfitting has been avoided in the model.The prediction results show that the model is stable and reliable, and the results are satisfying. |
| |
| Keywords: | Neural networks High-technical projects Investment Risk |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
