| 变结构人工神经网络模型及其在成矿预测中的应用 |
| |
| 引用本文： | 杨中宝,彭省临,王力.变结构人工神经网络模型及其在成矿预测中的应用[J].世界地质,2004,23(1):41-44,55. |
| |
| 作者姓名： | 杨中宝 彭省临 王力 |
| |
| 作者单位： | 中南大学,地学与环境工程学院,湖南,长沙,410083 |
| |
| 摘 要： | 针对人工神经网络成矿预测模型结构难以确定的问题，详细阐述了一种在模型训练中进行隐层数目及隐层单元数目动态调整的人工神经网络算法，并以VC 为开发工具实现了变结构人工神经网络成矿预测模型，经用华南26个岩体检验，回忆率及预测率均高达100％。该方法提供了一种面向具体问题的动态解决方案，在成矿预测工作中具有一定的实用性。
|
| 关 键 词： | 变结构人工神经网络 动态调整 成矿预测 VC＋＋ |
| 文章编号： | 1004-5589(2004)01-0041-05 |
Application of variable structure neural net model in metallogenic prognosis |
| |
| YANG Zhong-bao,PENG Sheng-lin,WANG Li.Application of variable structure neural net model in metallogenic prognosis[J].World Geology,2004,23(1):41-44,55. |
| |
| Authors: | YANG Zhong-bao PENG Sheng-lin WANG Li |
| |
| Abstract: | Aiming to solve the problem in metallogenic prognosis structure model of manual neural net, the authors discussed a sort of arithmetic for the dynamic creation of manual neural net to adjust the structure dynamically in the course of model building and introduced a variable structure manual neural net model for metallogenic prognosis by means of VC++ development. Tested by 26 rock bodies in South China,it indicates that both retrieve and prognosis rates are 100 percent. This method provides a dynamic way to resolve concrete problems.It's applicable in metallogenic prognosis. |
| |
| Keywords: | variable structure manual neural net dynamic creation metallogenic prognosis VC++ |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
