摘要
覆盖算法是张铃教授提出的一种构造性的神经网络学习算法,它根据样本数据自身的结构,构造性地建立了神经网络模型。这种构造性的神经网络模型与传统的前向神经网络相比具有运行速度快、精度高的特点。本文将构造性神经网络与基于商空间的时间序列结合起来建立模型,对煤矿瓦斯数据进行预测。实验结果表明该算法比传统的神经网络建模算法更适合于煤矿瓦斯预测。
Covering algorithm is proposed as a constructive neural network learning algorithm by professor Zhang Ling, which according to the structure of the sample data itself, constructive in establishing the neural network model. This structure of neural network model with the traditional feedforward neural networks compared to high speed operation and high accuracy. This paper will construct neural network and time series of space-based operators combine build models to predict the data of the coal mine. The experimental results show that the algorithm is more suitable for coal mine gas prediction than the traditional neural network modeling algorithm.
出处
《微计算机信息》
2010年第33期131-133,共3页
Control & Automation
基金
基金申请人:张月琴
项目名称:基于商空间的构造性数据挖掘方法的研究
基金颁发部门:山西省自然科学基金委(2008011028-1)
关键词
商空间
聚类分析
覆盖算法
时间序列
Commercial space
cluster analysis
covering algorithm
time series
