论文部分内容阅读
针对除铜过程影响工况的因素复杂、工况波动大等问题,提出了一种基于特征加权模糊支持向量机的工况评估方法.本文采用模糊隶属度对各个样本赋予不同的权值,对离群样本点与有效样本进行区分.根据信息增益计算各样本特征的权重,改进模糊支持向量机的核函数,进而构建基于特征权重的模糊支持向量机,实现除铜过程工况评估.分别利用标准实验数据集和现场数据进行测试,对比不同方法的分类精度,结果证明了本文工况评估方法的有效性和准确性.
Aiming at the problems such as complex working conditions of copper removal process and large fluctuation of working conditions, a condition-based evaluation method based on feature-weighted fuzzy support vector machine is proposed.In this paper, fuzzy membership is used to assign different weight to each sample, The outlier sample points are distinguished from valid samples, the weight of each sample feature is calculated according to the information gain, the kernel function of fuzzy support vector machine is improved, and then the fuzzy support vector machine based on feature weight is constructed to evaluate the condition of copper removal process. Standard experimental data sets and field data were tested to compare the classification accuracy of different methods. The results proved the validity and accuracy of the proposed method.