{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,10]],"date-time":"2026-10-10T13:55:58Z","timestamp":1791640558157,"version":"4.3.4"},"reference-count":29,"publisher":"Oxford University Press (OUP)","issue":"1","license":[{"start":{"date-parts":[[2008,1,4]],"date-time":"2008-01-04T00:00:00Z","timestamp":1199404800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/academic.oup.com\/journals\/pages\/open_access\/funder_policies\/chorus\/standard_publication_model"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2008,2,1]]},"abstract":"<jats:title>Summary<\/jats:title><jats:p>We suggest a technique, related to the concept of \u2018detection boundary\u2019 that was developed by Ingster and by Donoho and Jin, for comparing the theoretical performance of classifiers constructed from small training samples of very large vectors. The resulting \u2018classification boundaries\u2019 are obtained for a variety of distance-based methods, including the support vector machine, distance-weighted discrimination and kth-nearest-neighbour classifiers, for thresholded forms of those methods, and for techniques based on Donoho and Jin's higher criticism approach to signal detection. Assessed in these terms, standard distance-based methods are shown to be capable only of detecting differences between populations when those differences can be estimated consistently. However, the thresholded forms of distance-based classifiers can do better, and in particular can correctly classify data even when differences between distributions are only detectable, not estimable. Other methods, including higher criticism classifiers, can on occasion perform better still, but they tend to be more limited in scope, requiring substantially more information about the marginal distributions. Moreover, as tail weight becomes heavier the classification boundaries of methods designed for particular distribution types can converge to, and achieve, the boundary for thresholded nearest neighbour approaches. For example, although higher criticism has a lower classification boundary, and in this sense performs better, in the case of normal data, the boundaries are identical for exponentially distributed data when both sample sizes equal 1.<\/jats:p>","DOI":"10.1111\/j.1467-9868.2007.00631.x","type":"journal-article","created":{"date-parts":[[2010,8,19]],"date-time":"2010-08-19T16:34:57Z","timestamp":1282235697000},"page":"159-173","source":"Crossref","is-referenced-by-count":46,"title":["Theoretical Measures of Relative Performance of Classifiers for High Dimensional Data with Small Sample Sizes"],"prefix":"10.1093","volume":"70","author":[{"given":"Peter","family":"Hall","sequence":"first","affiliation":[{"name":"Australian National University , Canberra, and University of Melbourne , Australia"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yvonne","family":"Pittelkow","sequence":"additional","affiliation":[{"name":"Australian National University , Canberra , Australia"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Malay","family":"Ghosh","sequence":"additional","affiliation":[{"name":"University of Florida , Gainesville , USA"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"286","published-online":{"date-parts":[[2008,1,4]]},"reference":[{"key":"2023040708300349200_","first-page":"311","article-title":"Effect of high dimension: by an example of a two sample problem","volume":"6","author":"Bai","year":"1996","journal-title":"Statist. Sin."},{"key":"2023040708300349200_","doi-asserted-by":"crossref","first-page":"1643","DOI":"10.1214\/009117905000000233","article-title":"Phase transition of the largest eigenvalue for nonnull complex sample covariance matrices","volume":"33","author":"Baik","year":"2005","journal-title":"Ann. Probab."},{"key":"2023040708300349200_","doi-asserted-by":"crossref","first-page":"262","DOI":"10.1073\/pnas.97.1.262","article-title":"Knowledge-based analysis of microarray gene expression data by using support vector machines","volume":"97","author":"Brown","year":"2000","journal-title":"Proc. Natn. Acad. Sci. USA"},{"key":"2023040708300349200_","first-page":"955","article-title":"A tutorial on support vector machines for pattern recognition","volume":"2","author":"Burges","year":"1998","journal-title":"Data Minng Knowl. Disc."},{"key":"2023040708300349200_","doi-asserted-by":"crossref","first-page":"598","DOI":"10.1111\/j.1365-2966.2006.10381.x","article-title":"No higher criticism of the Bianchi corrected WMAP data","volume":"369","author":"Cayon","year":"2006","journal-title":"Mnthly Notes R. Astronom. Soc."},{"key":"2023040708300349200_","doi-asserted-by":"crossref","first-page":"826","DOI":"10.1111\/j.1365-2966.2005.09277.x","article-title":"Higher criticism statistic: detecting and identifying non-Gaussianity in the WMAP first year data","volume":"362","author":"Cayon","year":"2005","journal-title":"Mnthly Notes R. Astronom. Soc."},{"key":"2023040708300349200_","doi-asserted-by":"crossref","first-page":"202","DOI":"10.1109\/3477.990876","article-title":"The training of neural classifiers with condensed datasets","volume":"32","author":"Choi","year":"2002","journal-title":"IEEE Trans. Syst. Man Cybernet. B"},{"key":"2023040708300349200_","volume-title":"An Introduction to Support Vector Machines","author":"Cristianini","year":"2000"},{"key":"2023040708300349200_","doi-asserted-by":"crossref","first-page":"962","DOI":"10.1214\/009053604000000265","article-title":"Higher criticism for detecting sparse heterogeneous mixtures","volume":"32","author":"Donoho","year":"2004","journal-title":"Ann. Statist."},{"key":"2023040708300349200_","doi-asserted-by":"crossref","first-page":"96","DOI":"10.1198\/016214504000000089","article-title":"Large-scale simultaneous hypothesis testing: the choice of a null hypothesis","volume":"99","author":"Efron","year":"2004","journal-title":"J. Am. Statist. Ass."},{"key":"2023040708300349200_","first-page":"2681","article-title":"Recent results about the largest eigenvalue of random covariance matrices and statistical application","volume":"36","author":"El Karoui","year":"2005","journal-title":"Acta Phys. Polon. B"},{"key":"2023040708300349200_","doi-asserted-by":"crossref","first-page":"427","DOI":"10.1111\/j.1467-9868.2005.00510.x","article-title":"Geometric representation of high dimension, low sample size data","volume":"67","author":"Hall","year":"2005","journal-title":"J. R. Statist. Soc. B"},{"key":"2023040708300349200_","doi-asserted-by":"crossref","first-page":"349","DOI":"10.1137\/1107036","article-title":"Some limit theorems for stationary processes","volume":"4","author":"Ibragimov","year":"1962","journal-title":"Theory Probab. Appl."},{"key":"2023040708300349200_","first-page":"401","article-title":"Minimax detection of a signal for ln-balls","volume":"7","author":"Ingster","year":"1999","journal-title":"Math. Meth. Statist."},{"key":"2023040708300349200_","first-page":"395","article-title":"Adaptive detection of a signal of growing dimension: I","volume":"10","author":"Ingster","year":"2001","journal-title":"Math. Meth. Statist."},{"key":"2023040708300349200_","first-page":"37","article-title":"Adaptive detection of a signal of growing dimension: II","volume":"11","author":"Ingster","year":"2002","journal-title":"Math. Meth. Statist."},{"key":"2023040708300349200_","volume-title":"Manuscript","author":"Jin","year":"2002"},{"key":"2023040708300349200_","first-page":"255","volume-title":"A Festschrift to Honor Herman Rubin","author":"Jin","year":"2005"},{"key":"2023040708300349200_","volume-title":"Proc. PHYSTAT 2005: Statistical Problems in Particle Physics, Astrophysics and Cosmology","author":"Jin","year":"2006"},{"key":"2023040708300349200_","first-page":"2470","article-title":"Cosmological non-Gaussian signature detection: comparing performance of different statistical tests","volume":"15","author":"Jin","year":"2005","journal-title":"EURASIP J. Appl. Signal Process."},{"key":"2023040708300349200_","doi-asserted-by":"crossref","first-page":"295","DOI":"10.1214\/aos\/1009210544","article-title":"On the distribution of the largest principal component","volume":"29","author":"Johnstone","year":"2001","journal-title":"Ann. Statist."},{"key":"2023040708300349200_","article-title":"Distance weighted discrimination","author":"Marron","year":"2005","journal-title":"Manuscript"},{"key":"2023040708300349200_","first-page":"373","article-title":"Estimating the proportion of false null hypotheses among a large number of independently tested hypotheses","volume":"34","author":"Meinshausen","year":"2004","journal-title":"Ann. Statist."},{"key":"2023040708300349200_","volume-title":"Multidimensional Clustering Algorithms","author":"Murtagh","year":"1985"},{"key":"2023040708300349200_","doi-asserted-by":"crossref","first-page":"281","DOI":"10.1016\/S0304-4076(97)86569-4","article-title":"Subsampling for heteroskedastic time series","volume":"81","author":"Politis","year":"1997","journal-title":"J. Econometr."},{"key":"2023040708300349200_","doi-asserted-by":"crossref","first-page":"163","DOI":"10.1080\/10543409808835229","article-title":"The analysis of small-sample multivariate data","volume":"8","author":"Sarandasa","year":"1998","journal-title":"J. Biopharm. Statist."},{"key":"2023040708300349200_","doi-asserted-by":"crossref","DOI":"10.7551\/mitpress\/4175.001.0001","volume-title":"Learning with Kernels","author":"Sch\u00f6lokopf","year":"2001"},{"key":"2023040708300349200_","volume-title":"Estimation of Dependences based on Empirical Data","author":"Vapnik","year":"1982"},{"key":"2023040708300349200_","doi-asserted-by":"crossref","DOI":"10.1007\/978-1-4757-2440-0","volume-title":"The Nature of Statistical Learning Theory","author":"Vapnik","year":"1995"}],"container-title":["Journal of the Royal Statistical Society Series B: Statistical Methodology"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/api.wiley.com\/onlinelibrary\/tdm\/v1\/articles\/10.1111%2Fj.1467-9868.2007.00631.x","content-type":"unspecified","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/academic.oup.com\/jrsssb\/article-pdf\/70\/1\/159\/49796508\/jrsssb_70_1_159.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/jrsssb\/article-pdf\/70\/1\/159\/49796508\/jrsssb_70_1_159.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,2,24]],"date-time":"2025-02-24T20:29:56Z","timestamp":1740428996000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/jrsssb\/article\/70\/1\/159\/7109392"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2008,1,4]]},"references-count":29,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2008,1,4]]},"published-print":{"date-parts":[[2008,2,1]]}},"URL":"https:\/\/doi.org\/10.1111\/j.1467-9868.2007.00631.x","relation":{},"ISSN":["1369-7412","1467-9868"],"issn-type":[{"value":"1369-7412","type":"print"},{"value":"1467-9868","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2008,2]]},"published":{"date-parts":[[2008,1,4]]}}}