|
41. |
Regularized Gaussian Discriminant Analysis through Eigenvalue Decomposition |
|
Journal of the American Statistical Association,
Volume 91,
Issue 436,
1996,
Page 1743-1748
Halima Bensmail,
Gilles Celeux,
Preview
|
PDF (985KB)
|
|
摘要:
Friedman proposed a regularization technique (RDA) of discriminant analysis in the Gaussian framework. RDA uses two regularization parameters to design an intermediate classifier between the linear, the quadratic the nearest-means classifiers. In this article we propose an alternative approach, called EDDA, that is based on the reparameterization of the covariance matrix [Σk] of a groupGkin terms of its eigenvalue decomposition Σk= λkDkAkDk′, where λk specifies the volume of density contours ofGk, the diagonal matrix of eigenvalues specifies its shape the eigenvectors specify its orientation. Variations on constraints concerning volumes, shapes orientations λk,Ak, andDklead to 14 discrimination models of interest. For each model, we derived the normal theory maximum likelihood parameter estimates. Our approach consists of selecting a model by minimizing the sample-based estimate of future misclassification risk by cross-validation. Numerical experiments on simulated and real data show favorable behavior of this approach compared to RDA.
ISSN:0162-1459
DOI:10.1080/01621459.1996.10476746
出版商:Taylor & Francis Group
年代:1996
数据来源: Taylor
|
42. |
Book Reviews |
|
Journal of the American Statistical Association,
Volume 91,
Issue 436,
1996,
Page 1749-1754
Preview
|
PDF (1155KB)
|
|
摘要:
Statistical Analysis of Nonnormal DataJ. V. Deshpande, A. P. Gore, and A. Shanubhogue. New York: Wiley, 1995. viii + 240 pp. $40.95. Reviewed by Subha ChakrabortiUniversity of Alabama-Tuscaloosa
ISSN:0162-1459
DOI:10.1080/01621459.1996.10476747
出版商:Taylor & Francis Group
年代:1996
数据来源: Taylor
|
43. |
Telegraphic Reviews |
|
Journal of the American Statistical Association,
Volume 91,
Issue 436,
1996,
Page 1754-1755
Preview
|
PDF (286KB)
|
|
摘要:
Measure Theory and Probability (2nd Ed.)Malcolm Adams and Victor Guillemin. Cambridge, MA: Birkhauser Boston, 1995. xiv + 205 pp. $26.50. Reviewed byRL
ISSN:0162-1459
DOI:10.1080/01621459.1996.10476748
出版商:Taylor & Francis Group
年代:1996
数据来源: Taylor
|
44. |
A General Representation of Equally Correlated Variates |
|
Journal of the American Statistical Association,
Volume 91,
Issue 436,
1996,
Page 1756-1756
HerbertA. David,
Preview
|
PDF (66KB)
|
|
ISSN:0162-1459
DOI:10.1080/01621459.1996.10476749
出版商:Taylor & Francis Group
年代:1996
数据来源: Taylor
|
45. |
Addendum and Corrections |
|
Journal of the American Statistical Association,
Volume 91,
Issue 436,
1996,
Page 1757-1757
Preview
|
PDF (183KB)
|
|
ISSN:0162-1459
DOI:10.1080/01621459.1996.10476750
出版商:Taylor & Francis Group
年代:1996
数据来源: Taylor
|
46. |
Editorial Board Page |
|
Journal of the American Statistical Association,
Volume 91,
Issue 436,
1996,
Page -
Preview
|
PDF (200KB)
|
|
摘要:
This article has no abstract
ISSN:0162-1459
DOI:10.1080/01621459.1996.10476705
出版商:Taylor & Francis Group
年代:1996
数据来源: Taylor
|
|