论文部分内容阅读
本文采用模式识别方法推断烧结矿质量。在给出模糊系统聚类分析算法基础上,用软件实现了基于模糊聚类分类器和动态聚类分类器,并用现场实测的样本采用“留一法”分别对这两种分类器性能进行检验。结果表明:模糊聚类分析法对于先验知识较少、样本量不大时,性能较佳。
In this paper, pattern recognition method to infer the quality of sinter. Based on the fuzzy clustering analysis algorithm, the fuzzy clustering classifier and dynamic clustering classifier are implemented by software, and the performance of these two classifiers are tested respectively by the “one-leave-only method” . The results show that fuzzy clustering analysis has better performance for less prior knowledge and less sample size.