基于Matlab的自组织神经网络在油气层识别中的应用研究

介绍了自组织竞争网络和自组织影射网络的原理,对自组织竞争网络和自组织影射网络的优缺点进行了比较。采用大庆的油气层数据建立网络模型,对网络结构的参数进行了优化并对输入样本进行了聚类分析。数据分析表明自组织竞争网络和自组织影射网络都有较好的聚类结果,自组织竞争网络较自组织影射网络方法识别出的结果更客观可靠,是油气层识别的一种有效方法。

Oil & Gas Formation Identification Based on Self-Organization Neural Network of Matlab

The theory of self-organization neural network is introduced. The advantage and disadvantage of self-organization competitive neural network and self-organization mapping neural network are compared. we optimize the parameter of net structure and make the cluster analysis of input sample. Oil-gas data of Daqing is used to build network model. The parameters of network structure are optimized, and the cluster analysis of input sample is made. The result of data analysis shows that the method of self-organization competitive neural network and self-organization mapping neural network is objective and reliability. This network is proved to be an effective method to improve ability of oil & gas formation identification.