摘要
本文依据分形理论[1]和支持向量机[2],建立矿井瓦斯涌出量的分形支持向量机预测模型。首先对瓦斯涌出量数据进行时间序列分形处理,进行相空间重构(即求嵌入空间的维数),然后将嵌入维数作为支持向量机的输入节点数,采用支持向量机对矿井瓦斯涌出量作出预测。结果表明:预测精度高达99.9%,说明将此模型用于瓦斯涌出量的预测是可行的。
In this paper, We present a model of forecasting of Gas Emission based on Fractal theory and Support Vector Machine. We first process the training data to get the space dimension which is used as the amount of SVM's node, then we make the forecasting based on SVM. The results show that the precision has come up to 99.999% ,which presents the model is reliable.
出处
《中国科技论文》
CAS
2006年第3期203-207,共5页
China Sciencepaper
关键词
分形
支持向量机
时间序列
预测
fractal
support vector machine
time series
forecasting
