地学信息挖掘中RPROP和分层动量增项自适应BP算法应用研究  

Study on Stratified Momentum Term Adaptive BP Algorithm in Multi-geoobject Information Mining

作　　者：刘伟[1] 徐勇[2] 

机构地区：[1]吉林建筑工程学院基础科学部,吉林长春130021 [2]吉林建筑工程学院计算机科学与工程学院,吉林长春130021

出　　处：《数学的实践与认识》2009年第8期112-117,共6页Mathematics in Practice and Theory

基　　金：国家高技术研究发展计划(863计划)(2006AA06Z114)

摘　　要：采用RPROP和分层动量增项自适应BP算法,从最小误差、收敛速度和运算次数方面对地球化学信息进行了研究,并进行了比较分析;针对本区域的样本数据训练结果,确定采用分层动量自适应算法进行后续预测统计工作,为矿区靶区预测提供支持.RPROP and Stratified momentum term adaptive BP algorithm is used, it has been studied from the smallest error, the convergence of computing speed and frequency of geochemical information, and has been comparatived analysis; to the sample of the results of training data, make sure to adopt the stratified momentum term adaptive BP algorithm proccess the Census and Statistics forecast, and the study can serve as the support of the target forecast of the mining areas.

关 键 词：分层增量自适应 RPROP BP算法 梯度 

分 类 号：P628.2[天文地球—地质矿产勘探]