摘要
简要回顾了人工神经网络的发展历史,介绍了信息计量学的概念和人工神经网络算法在信息计量学领域中的应用,警示了在人工神经网络应用中的"过拟合"现象。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线性神经网络模型,模型避免了"过拟合"现象的发生。通过对有关数据进行实例计算,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。
The development history of artificial neural network, the conception of informetrics and the application of neural network algorithms to the research fields of informetrics are introduced,and the problem of overfitting in the application of neural network methods is pointed out. Based on the analysis of the risk factors in investing in high tech projects, a linear neural network model is presented to predict the investment risks, and the problem of overfitting is avoided in the model. The calculation of data taken from the related literature is carried out,and the prediction results show that the model is stable and reliable and the results are satisfying.
出处
《合肥工业大学学报（自然科学版）》
CAS
CSCD
2004年第7期851-854,共4页
Journal of Hefei University of Technology：Natural Science
基金
安庆师范学院科研基金资助项目(2003YWY025)
关键词
神经网络
过拟合
信息计量学
投资风险评价
neural network
overfitting
informetrics
investment risk evaluation
