From interval arithmetic to multi-peak/fuzzy distribution arithmetic-comprehensive numerical analysis for uncertain systems-

点击翻译按钮获取中文摘要

In this research, conventional clustering methods for uncertain data are reviewed from the probabilistic point of view, new clustering methods are proposed for clustering, PCA, or linear clustering in the data space or in the feature space, which are obtained by solving the optimization problem adding squared Euclidean norm of the value of uncertainty for data to the original objective functions.