Principal curves are smooth one-dimensional curves that pass through the middle of a p-dimensional data set, providing a nonlinear summary of the data. They are nonparametric, and their shape is suggested by the data. The algorithm for constructing principal curves starts with some prior summary, such as the usual principal-component line. The curve in each successive iteration is a smooth or local average of the p-dimensional points, where the definition of local is based on the distance in arc length of the projections of the points onto the curve found in the previous iteration. In this article principal curves are defined, an algorithm for their construction is given, some theoretical results are presented, and the procedure is compared to other generalizations of principal components. Two applications illustrate the use of principal curves. The first describes how the principal-curve procedure was used to align the magnets of the Stanford linear collider. The collider uses about 950 magnets in a roughly circular arrangement to bend electron and positron beams and bring them to collision. After construction, it was found that some of the magnets had ended up significantly out of place. As a result, the beams had to be bent too sharply and could not be focused. The engineers realized that the magnets did not have to be moved to their originally planned locations, but rather to a sufficiently smooth arc through the middle of the existing positions. This arc was found using the principal-curve procedure. In the second application, two different assays for gold content in several samples of computer-chip waste appear to show some systematic differences that are blurred by measurement error. The classical approach using linear errors in variables regression can detect systematic linear differences but is not able to account for nonlinearities. When the first linear principal component is replaced with a principal component is replaced with a principal curve, a local "bump" is revealed, and bootstrapping is used to verify its presence.
The Journal of the American Statistical Association (JASA) has long been considered the premier journal of statistical science. Science Citation Index reported JASA was the most highly cited journal in the mathematical sciences in 1991-2001, with 16,457 citations, more than 50% more than the next most highly cited journals. Articles in JASA focus on statistical applications, theory, and methods in economic, social, physical, engineering, and health sciences and on new methods of statistical education.
Building on two centuries' experience, Taylor & Francis has grown rapidlyover the last two decades to become a leading international academic publisher.The Group publishes over 800 journals and over 1,800 new books each year, coveringa wide variety of subject areas and incorporating the journal imprints of Routledge,Carfax, Spon Press, Psychology Press, Martin Dunitz, and Taylor & Francis.Taylor & Francis is fully committed to the publication and dissemination of scholarly information of the highest quality, and today this remains the primary goal.
This item is part of JSTOR collection
For terms and use, please refer to our Terms and Conditions
Journal of the American Statistical Association
© 1989 American Statistical Association
Request Permissions
