论文部分内容阅读
将序贯数论优化（ＳＮＴＯ）法用于聚类分析的优化过程，同时根据分类结果，基于样本之间的相似性和差异性，构造一Ｒｏｕｓｅｕｗ定义的函数并做侧影图，用以评价分类效果。结果表明，ＳＮＴＯ法原理清晰且易于实现，是一种比较实用的全局优化方法。将构造的侧影图用于聚类分析结果的解释，特别是分类数的确定也基本令人满意。
The sequential number theory optimization (SNTO) method is applied to the optimization of clustering analysis. Based on the classification results, a Rouseuw-defined function is constructed based on the similarity and difference between the samples and the silhouette map is used to evaluate the classification effect . The results show that the principle of SNTO method is clear and easy to implement, which is a practical global optimization method. The constructed silhouette map is used to explain the clustering analysis results, especially the classification number is also basically satisfactory.