面向材料基因工程的人工神经网络研究

Research on Artificial Neural Network for Material Genetic Engineering

介绍了人工神经网络的基本原理,概述了三种典型人工神经网络模型的特点及其在材料基因工程领域中的应用范围。当实验数据充足,可选择RBF人工神经网络;当需要提供较强的理论依据以及实验数据少的情况下,BP神经网络是比较优先的选择;当遇到分类与识别问题时可应用自组织竞争神经网络。

The basic principles of artificial neural network （ANN） were reviewed. The characteristics of the three typical ANN models and its application range in the fields of material genetic engineering were briefly illustrated. The RBF artificial neural network could be selected when the experimental data was sufficient. BP neural network was the preferred choice when it was necessary to provide strong academic basis and less experimental data. Self-organizing competition neural networks could be applied when encountering the classification and identification problems.