神经网络环境下矿山规模的预测  

The Prediction of Mine Scale Under Nerve Network Environment

作　　者：雷万杉[1] 王忠文[2] 范继璋[1] 

机构地区：[1]吉林大学综合信息矿产预测研究所,吉林长春130026 [2]吉林工程技术师范学院基础科学系,吉林长春130052

出　　处：《吉林工程技术师范学院学报》2008年第4期57-60,共4页Journal of Jilin Engineering Normal University

摘　　要：矿床规模影响因子较多,并且这些因子与矿床规模间存在着非常复杂的非线性关系,一直以来只能实现定性化预测,预测方法一般采用逻辑信息法、回归分析方法和蒙特卡罗法来进行,这些方法计算复杂,结果不稳定,为预测的进行和结果的解释带来很多的困扰。BP神经网络可以实现从输入到输出的高度非线性映射。本文对矿床资料进行规范化处理,并进行适当插值,利用MATLAB工具箱成功实现了对待评矿床的定量化预测。The influence factor of ore deposit scale more, and these factors and the ore deposit scale have the extremely complex non -linear relations, since continuously only has been able to realize the qualitative forecast, the method of forecast generally uses the logical information law, the regression analysis method and the Monte Carlo method carries on, these method conlputations is complex, the result is unstable, brings very many puzzles for forecast carrying on with the result explanation. The BP nerve network may achieve the non -linear mapping from inputs to the output. This article carries on standardized processing to the ore deposit material, and carries on the suitable interpolation,realized the treatment using the MATLAB toolbox success to comment theore deposit quantitative forecast, has obtained the good appraisal result, to enhanced the mine appraisal the precision and objective has the extremely positive significance.

关 键 词：神经网络 综合信息矿产预测 矿山规模 

分 类 号：P618.51[天文地球—矿床学]