摘要
自组织映射网络是一种学习速度很快的神经网络,可以用于分类、聚类、解释等问题。本文用该方法对两批数据进行了油气检测试验。试验表明该方法有如下特点:(1)网络输出层越大,学习能力越强;(2)学习的结果好坏还依赖于样本的选择;(3)可以得到很高的分类正确率。自组织映射方法值得在地球物理勘探及地质分类问题中应用。
Self-organizing mapping network is a fast learning neural network used to deal with problems of classification, clustcring, interpretation and so on. In the paper, the method is tested with an experiment for hydrocarbon detection on two data sets. The results show: (1)the biger the output layer of network, the greater the learning ability, (2) the learning ability still depends on the selection for samples, and (3)very good results could be reached while the method is in application to classification. This method is worth recommending in geophysical prospecting and geological classification applications.
出处
《石油物探》
EI
CSCD
北大核心
1994年第4期56-64,共9页
Geophysical Prospecting For Petroleum
基金
中国科学院基金
国家自然科学基金
大庆石油管理局
中国石油天然气总公司联合资助
关键词
油气检测
神经网络
地球物理勘探
Self-Organizing Mapping,Neural Network,Hydrocarbon Detection,Classification