STRATEGIES OF SELECTING THE BASIS VECTOR SET IN THE RELATIVE MDS
研究了一种结合聚类与多维尺度分析的大规模多维数据可视化方法，提出基向量选择策略的改进，实验表明新方法在可视化质量和计算开销上优于原算法。
In this paper, a method of large multidimensional data visualization that associates the multidimensional scaling (MDS) with clustering is modified and investigated. In the original algorithm, the visualization process is divided into three steps: the basis vector set is constructed using the k‐means clustering method; this set is projected onto the plane using the MDS algorithm; the remaining data set is visualized using the relative MDS algorithm. We propose a modification which differs from the original algorithm in the strategy of selecting the basis vectors. In our modification, the set of basis vectors consists of vectors that are selected from k clusters in a new way. The experimental investigation showed that the modification exceeds the original algorithm in visualization quality and computational expenses.