我国各省区煤炭生产数值分布的自组织分类  

Self-organizing Classification for the Distribution of the Content of Provinces and Autonomous Regions Coal Production Data Based on SOM Neural Networks

作　　者：杨纪青[1] 

机构地区：[1]襄樊学院数计学院,湖北襄樊441053

出　　处：《广州化工》2009年第9期11-12,共2页GuangZhou Chemical Industry

基　　金：湖北省"十一五"教育科学发展规划项目资助(课题号:2006B131)

摘　　要：为实现煤炭生产数值分布的自组织分类,用SOM神经网络技术,以《中国工业统计年鉴》提供的我国各省区煤炭生产数据为学习和测试样本,训练和检测神经网络对我国各省区煤炭生产数值分布的自组织分类。实验显示:经过1000步的训练,SOM神经网络矩阵映射重复操作归类相同率,在剔除特异结果后达到90%以上。这一结果证实,利用SOM神经网络技术,可以对我国各省区煤炭生产数据参数组合取值分布进行客观分类。The serf - organizing classification for the distribution of content of numerical distribution of coal production was based on SOM neural networks. Using numerical distribution of coal production as study and test samples provided by province, the self - organizing classification for the distribution of content of lead and arsenic in the numerical distribution of coal production was trained and tested. After 1000 step by step training, the experiments showed that the precision of SOM neural networks matrix mapping was 10%, and the same rate of classifying repeat operations was 90%. The results explained that an objective classification for the distribution of content of lead and arsenic in the numerical distribution of coal production using the SOM neural networks technology could be made, and which could be provided as basis for the rating of numerical distribution of coal production quality, properties forecast, defects diagnosis and the recognition of type.

关 键 词：各省区煤炭生产数值 自组织分类 SOM神经网络 煤炭生产数值分布 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] F426.21[自动化与计算机技术—计算机科学与技术]