摘要
目的根据无线通信建筑环境干扰的特征变化,寻找确定通信控制终端发射功率分类的方法.方法采用粗糙集软计算特征约简方法对通信控制终端的原始特征进行优化,去除冗余特征,并采用BP网络对测试样本中随机选取的样本进行测试.结果通过对该方法仿真,证明了该方法具有良好的识别精度.结论根据通信控制终端的特征,提出基于粗糙集软计算的特征约简方法,该方法能够提高神经网络识别分类的可靠性和快速性.
The paper aims to determine the classification method of characteristic optimization communication controlling terminal, based on the change of the characteristics of wireless internet building environment disturbance. We optimized the primitive characteristic of communication controlling terminal through rough sets software computation characteristic reduction method, eliminated redundant characteristic and used the BP network to test the stochastic samples. The result shows that this method is better than the traditional one in recognizing precision through simulation, The reduction method proposed here can enhance reliability and the rapidity of the neural network recognition classification.
出处
《沈阳建筑大学学报（自然科学版）》
CAS
2008年第1期157-160,共4页
Journal of Shenyang Jianzhu University：Natural Science
基金
辽宁省教育厅科学技术研究项目(20060700)
关键词
粗糙集
BP神经网络
发射功率
特征提取
分类器
rough sets
BP neural network
launch power
feature optimization
classifier
