摘要
针对现有煤泥浮选加药量预测精度不足的现状,提出了基于广义回归神经网络(GRNN)的浮选加药量预测。首先介绍了GRNN的网络结构,然后通过交叉验证搜索算法确定了模型的结构参数,最后通过与BP网络模型的比较,得出了该模型在算法时间和预测精度方面的优越性,更加适应于浮选加药的预测。
In view of status quo that the existing coal slime flotation dosage prediction precision is deficient, the prediction of flotation dosage based on general regression neural network （GRNN） is presented. Firstly, the network structure of GRNN is introduced, and then the structural parameters of the model are determined through the cross validation search algorithm, finally obtains that this model is superior to the model of BP network in time and prediction accuracy by comparing with the model BP network, and is more suitable to forecast the floatation dosing.
出处
《煤炭技术》
CAS
北大核心
2017年第2期286-288,共3页
Coal Technology
关键词
浮选加药
GRNN
交叉搜索
预测
flotation medicine
GRNN
cross validation
prediction
