Facial Representation of Multivariate Data
介绍了Chernoff开发的一种将多维空间中的数据点表示为卡通人脸的方法，说明了其构造机制和几何关系，并展示了该方法在检测变量间微妙但重要关系中的应用。
An application of a relatively new method of depicting multivariate data graphically, this article presents a method developed by Chernoff consisting of representing a point in k-dimensional space as a cartoon of a face. The authors briefly describe the mechanics of constructing a face as well as the geometric relationships involved, illustrate the use of faces as a method of portraying multivariate data graphically, and exemplify the use of this method in detecting subtle, yet potentially significant relationships among variables.