数据开采的BP神经网络方法及其应用

DATA MINING METHOD OF BP ARTIFICIAL NEURAL NETWORK AND ITS APPLICATION

简要介绍了数据开采的基本概念和地学意义，并运用人工神经网络中的前向多层反传学习算法（ＢＰ），在导师信号监督下对网络进行训练学习，分别用岩石化学成分和有利找矿标志作为导师知识引导知识发现，实现火成岩目标信息挖掘、模式识别分类和金矿找矿预测．

The basic concept of data mining and its geological meaning are expounded. Also introduced are a data mining method of back-propagation artificial neural network and its learning algorithm, and application to the KDD and gold deposit forecast. BP neural network consists of input, hidden and output layers. The details of the layers of the examples are: (1) input layers are the chemical composition of magmatite, and output layers, the discern parameters of rock character, their application to the KDD and rock classification; (2) input layers are the prospecting indicator of structures, beds, altered rocks and gold anomaly ( w (Au)>25×10 -9 ), while the output is mineralization prevailing level to optimization of gold exploration target. Both examples have achieved good result.