{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,12,29]],"date-time":"2025-12-29T02:16:13Z","timestamp":1766974573277,"version":"3.48.0"},"reference-count":59,"publisher":"Wiley","issue":"6","license":[{"start":{"date-parts":[[2025,12,9]],"date-time":"2025-12-09T00:00:00Z","timestamp":1765238400000},"content-version":"vor","delay-in-days":8,"URL":"http:\/\/onlinelibrary.wiley.com\/termsAndConditions#vor"}],"content-domain":{"domain":["onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["Numerical Linear Algebra App"],"published-print":{"date-parts":[[2025,12]]},"abstract":"<jats:title>ABSTRACT<\/jats:title>\n                  <jats:p>Linear discriminant analysis (LDA) is a classical method for feature extraction and dimensionality reduction. Mathematically, the LDA method resorts to a trace\u2010ratio problem; however, it often suffers from the small\u2010sample\u2010size (SSS) problem for data with high dimension. Matrix exponential discriminant analysis (EDA) methods are efficient ways to deal with the SSS problem. To the best of our knowledge, however, all the existing matrix exponential\u2010based discriminant methods reduce to ratio\u2010trace problems, rather than trace\u2010ratio problems. This goes against the motivation of the original Fisher criterion for LDA. Indeed, it has been shown that the trace\u2010ratio model can yield markedly improved recognition results for supervised learning tasks. Therefore, it is interesting to investigate discriminant analysis methods that can combine the matrix exponential and trace\u2010ratio model. To fill in this gap, we first propose a matrix exponential trace\u2010ratio model and present a fast iterative exponential trace\u2010ratio (FIETR) algorithm for solving it efficiently. The key is to equivalently transform a large\u2010scale matrix exponential computation problem into a much smaller one. Second, we give the convergence of the iterative exponential trace\u2010ratio algorithm, and show that it can converge faster than the popular iterative trace\u2010ratio algorithm (ITR), which is the state\u2010of\u2010the\u2010art algorithm for the trace\u2010ratio problem. Thanks to the property of matrix exponential, we prove that the exponential trace\u2010ratio values generated by our algorithm converge locally and (at least) quadratically to the global optimal objective value. Third, to accelerate the FIETR method further, we propose a non\u2010iterative algorithm for solving the matrix exponential trace\u2010ratio problem efficiently. Numerical experiments on some real\u2010world data sets illustrate the numerical behavior of the proposed algorithms.<\/jats:p>","DOI":"10.1002\/nla.70053","type":"journal-article","created":{"date-parts":[[2025,12,10]],"date-time":"2025-12-10T00:40:55Z","timestamp":1765327255000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Exponential Trace\u2010Ratio Method for High Dimensionality Reduction"],"prefix":"10.1002","volume":"32","author":[{"given":"Qi\u2010Yuan","family":"Zhang","sequence":"first","affiliation":[{"name":"School of Mathematics China University of Mining and Technology  Xuzhou P.R. China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Gang","family":"Wu","sequence":"additional","affiliation":[{"name":"School of Mathematics China University of Mining and Technology  Xuzhou P.R. China"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"311","published-online":{"date-parts":[[2025,12,9]]},"reference":[{"key":"e_1_2_11_2_1","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2022.3194896"},{"key":"e_1_2_11_3_1","first-page":"846","volume-title":"IEEE Computer Society Conference on Computer Vision and Pattern Recognition","author":"Chen H.","year":"2005"},{"key":"e_1_2_11_4_1","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2019.2919597"},{"volume-title":"Introduction to Statistical Pattern Recognition","year":"1999","author":"Fukunaga K.","key":"e_1_2_11_5_1"},{"key":"e_1_2_11_6_1","first-page":"153","article-title":"Locality Preserving Projections","volume":"16","author":"He X.","year":"2004","journal-title":"Advances in Neural Information Processing Systems"},{"key":"e_1_2_11_7_1","doi-asserted-by":"publisher","DOI":"10.1002\/nla.743"},{"key":"e_1_2_11_8_1","first-page":"5322","article-title":"Sparse PCA via l2,p$$ {\\mathrm{l}}_{2,\\mathrm{p}} $$\u2010Norm Regularization for Unsupervised Feature Selection","volume":"45","author":"Li Z.","year":"2025","journal-title":"IEEE Transactions on Pattern Analysis and Machine Intelligence"},{"volume-title":"Foundations of Machine Learning","year":"2012","author":"Mohri M.","key":"e_1_2_11_9_1"},{"key":"e_1_2_11_10_1","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-642-27497-8"},{"key":"e_1_2_11_11_1","doi-asserted-by":"publisher","DOI":"10.1002\/0470854774"},{"key":"e_1_2_11_12_1","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2007.250598"},{"key":"e_1_2_11_13_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.imavis.2005.11.006"},{"key":"e_1_2_11_14_1","doi-asserted-by":"publisher","DOI":"10.1109\/TKDE.2018.2842023"},{"key":"e_1_2_11_15_1","doi-asserted-by":"publisher","DOI":"10.1007\/978-1-4757-1904-8"},{"key":"e_1_2_11_16_1","doi-asserted-by":"publisher","DOI":"10.1162\/jocn.1991.3.1.71"},{"key":"e_1_2_11_17_1","doi-asserted-by":"publisher","DOI":"10.1109\/34.598228"},{"key":"e_1_2_11_18_1","doi-asserted-by":"publisher","DOI":"10.1111\/j.1469-1809.1936.tb02137.x"},{"key":"e_1_2_11_19_1","doi-asserted-by":"publisher","DOI":"10.1137\/120864799"},{"key":"e_1_2_11_20_1","doi-asserted-by":"publisher","DOI":"10.1007\/s10994-020-05937-w"},{"key":"e_1_2_11_21_1","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2019.2901309"},{"key":"e_1_2_11_22_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.patcog.2007.07.022"},{"key":"e_1_2_11_23_1","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2019.2904701"},{"key":"e_1_2_11_24_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.cognition.2017.12.005"},{"key":"e_1_2_11_25_1","doi-asserted-by":"publisher","DOI":"10.1080\/01621459.1989.10478752"},{"key":"e_1_2_11_26_1","doi-asserted-by":"publisher","DOI":"10.2307\/2986198"},{"key":"e_1_2_11_27_1","doi-asserted-by":"publisher","DOI":"10.1137\/110851377"},{"key":"e_1_2_11_28_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.patrec.2011.11.028"},{"key":"e_1_2_11_29_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.patcog.2009.10.004"},{"key":"e_1_2_11_30_1","doi-asserted-by":"publisher","DOI":"10.1109\/TSMCB.2009.2024759"},{"key":"e_1_2_11_31_1","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2024.3362512"},{"key":"e_1_2_11_32_1","doi-asserted-by":"publisher","DOI":"10.1109\/TCYB.2020.3025205"},{"key":"e_1_2_11_33_1","doi-asserted-by":"publisher","DOI":"10.1109\/TCSVT.2018.2799214"},{"key":"e_1_2_11_34_1","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2021.3133351"},{"key":"e_1_2_11_35_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ins.2024.121444"},{"key":"e_1_2_11_36_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.knosys.2020.105630"},{"key":"e_1_2_11_37_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.csda.2024.108108"},{"key":"e_1_2_11_38_1","doi-asserted-by":"publisher","DOI":"10.1002\/nla.2594"},{"key":"e_1_2_11_39_1","first-page":"5568","article-title":"A Novel Formulation of Trace Ratio Linear Discriminant Analysis","volume":"33","author":"Wang J.","year":"2022","journal-title":"IEEE Transactions on Pattern Analysis and Machine Intelligence"},{"key":"e_1_2_11_40_1","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2024.3355422"},{"key":"e_1_2_11_41_1","doi-asserted-by":"publisher","DOI":"10.1137\/1.9780898717778"},{"key":"e_1_2_11_42_1","first-page":"2196","article-title":"Marginal Fisher Analysis Based on Matrix Exponential Transformation","volume":"10","author":"He J.","year":"2014","journal-title":"Chinese Journal of Computers"},{"key":"e_1_2_11_43_1","doi-asserted-by":"publisher","DOI":"10.1109\/TIP.2013.2297020"},{"key":"e_1_2_11_44_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.neucom.2011.07.007"},{"key":"e_1_2_11_45_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.patcog.2016.07.029"},{"key":"e_1_2_11_46_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.patcog.2020.107530"},{"key":"e_1_2_11_47_1","doi-asserted-by":"publisher","DOI":"10.1016\/S0031-3203(99)00139-9"},{"key":"e_1_2_11_48_1","doi-asserted-by":"publisher","DOI":"10.4208\/csiam-am.2021.nla.03"},{"key":"e_1_2_11_49_1","first-page":"1","article-title":"Trace\u2010Ratio vs. Ratio\u2010Trace for Dimensionality Reduction","author":"Wang H.","year":"2007","journal-title":"IEEE Computer Society Conference on Compute Vision and Pattern Recognition"},{"key":"e_1_2_11_50_1","doi-asserted-by":"publisher","DOI":"10.1137\/080720863"},{"key":"e_1_2_11_51_1","doi-asserted-by":"publisher","DOI":"10.1109\/TNN.2009.2015760"},{"key":"e_1_2_11_52_1","doi-asserted-by":"publisher","DOI":"10.1007\/s10092-020-00366-6"},{"key":"e_1_2_11_53_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.patcog.2016.08.020"},{"key":"e_1_2_11_54_1","doi-asserted-by":"publisher","DOI":"10.1137\/0707001"},{"volume-title":"Matrix Perturbation Analysis (In Chinese)","year":"2001","author":"Sun G.","key":"e_1_2_11_55_1"},{"volume-title":"Numerical Analysis","year":"2016","author":"Burden R.","key":"e_1_2_11_56_1"},{"key":"e_1_2_11_57_1","doi-asserted-by":"publisher","DOI":"10.1137\/S0895479895290954"},{"key":"e_1_2_11_58_1","doi-asserted-by":"publisher","DOI":"10.1007\/s10957-011-9832-4"},{"key":"e_1_2_11_59_1","doi-asserted-by":"publisher","DOI":"10.1109\/TIT.1967.1053964"},{"key":"e_1_2_11_60_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.apnum.2020.09.013"}],"container-title":["Numerical Linear Algebra with Applications"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/nla.70053","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,12,29]],"date-time":"2025-12-29T02:13:02Z","timestamp":1766974382000},"score":1,"resource":{"primary":{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/10.1002\/nla.70053"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,12]]},"references-count":59,"journal-issue":{"issue":"6","published-print":{"date-parts":[[2025,12]]}},"alternative-id":["10.1002\/nla.70053"],"URL":"https:\/\/doi.org\/10.1002\/nla.70053","archive":["Portico"],"relation":{},"ISSN":["1070-5325","1099-1506"],"issn-type":[{"type":"print","value":"1070-5325"},{"type":"electronic","value":"1099-1506"}],"subject":[],"published":{"date-parts":[[2025,12]]},"assertion":[{"value":"2025-04-16","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-11-24","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-12-09","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e70053"}}