Detail New methodology in regression and multivariate quality control via data depth Bibliografi Author: Teng, Huey-Chung ; Liu, Regina (Advisor); Singh, Kesar (Advisor) Topik: STATISTICS Bahasa: (EN )    ISBN: 0-599-61701-2     Penerbit: RUTGERS THE STATE UNIVERSITY OF NEW JERSEY - NEW BRUNSWICK     Tahun Terbit: 2000     Jenis: Theses - Dissertation Fulltext: 9958417.pdf (0.0B; 2 download) Abstract Several depth-based nonparametric multivariate methods are developed for regression and statistical quality control charts. The regression methods include the pick-p approach, the simplicial-intercept approach and the depth-weighted approach. They are shown to be robust in the presence of outliers. Some real datasets with extreme outliers are presented, and outcomes are quite supportive of our methods. As a byproduct, the objective function itself in the simplicial-intercept approach can provide a natural measure of goodness-of-fit of the resulting linear regression. In the realm of multivariate statistical quality control, the so-called r- and Q-charts have been constructed based on the idea of reducing each multivariate characteristic to a univariate index, namely its relative center-outward rank induced by a data depth. In this dissertation, we develop further the r- and Q-charts by applying various moving average schemes to the ranks. The moving average scheme allows our new charts to achieve a superior capability in detecting location changes in a process. This superiority is particularly acute in the non-elliptical processes. Graphical results based on many datasets simulated from different distributions are presented to highlight this property. Opini Anda Klik untuk menuliskan opini Anda tentang koleksi ini! Lihat Sejarah Pengadaan  Konversi Metadata   Kembali

Copyright © 2006, 2007 Unika Atma Jaya, all rights reserved