高维数据的2D图单点表示原理

2D graphical representation reflecting high-dimensional observation with sole point

高维数据的单点2D图表示是2D图表示的重要分支。由于其可在单幅多元图中显示多个观察,因此适用于模式识别领域中的特征选择与分类空间形成。本文根据多元图中单点映射的变量数目的不同对单点图表示进行了对比,并分析了其各自的适用范围与优缺点。

2D graphical representation which reflects high-dimensional observation with sole point is one important branch of 2D graphical representations. The representation is competent for feature selection and class space building as it can display a set of observations in one graph. In this paper, the typical representations are analyzed according to the relationship between the sole point and number of attributes. The advantages, disadvantages of each representation and its applicability are also discussed.