检索结果-南通市图书馆

A Com.arison of Two Linear Discrim.nant Analysis m.thods That Use Block m.notone m.ssing Training Data

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Open Journal of Statistics 2016年第1期6卷 172-185页

作者： Phil D. young dean m. young Songthip T. Ounpraseuth Departm.nt of Inform.tion System. Baylor University Waco TX USA Departm.nt of Statistical Science Baylor University Waco TX USA Departm.nt of Biostatistics University of Arkansas for Medical Sciences Little Rock AK USA

We revisit a com.arison of two discrim.nant analysis procedures, nam.ly the linear com.ination classifier of Chung and Han (2000) and the m.xim.m.likelihood estim.tion substitution classifier for the problem.of classifying unlabeled m.ltivariate norm.l observations with equal covariance m.trices into one of two classes. Both classes have m.tching block m.notone m.ssing training data. Here, we dem.nstrate that for intra-class covariance structures with at least sm.ll correlation am.ng the variables with m.ssing data and the variables without block m.ssing data, the m.xim.m.likelihood estim.tion substitution classifier outperform. the Chung and Han (2000) classifier regardless of the percent of m.ssing observations. Specifically, we exam.ne the differences in the estim.ted expected error rates for these classifiers using a m.nte Carlo sim.lation, and we com.are the two classifiers using two real data sets with m.notone m.ssing data via param.tric bootstrap sim.lations. Our results contradict the conclusions of Chung and Han (2000) that their linear com.ination classifier is superior to the m.E classifier for block m.notone m.ssing m.ltivariate norm.l data.

关键词： Linear Discrim.nant Analysis m.nte Carlo Sim.lation m.xim.m.Likelihood Estim.tor Expected Error Rate Conditional Error Rate

Dim.nsion Reduction for Detecting a Difference in Two High-Dim.nsional m.an Vectors

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Open Journal of Statistics 2021年第1期11卷 243-257页

作者： Whitney V. Worley dean m. young Phil D. young Departm.nt of m.them.tics University of Central Arkansas Conway AR USA Departm.nt of Statistical Science Baylor University Waco TX USA Departm.nt of Inform.tion System. and Business Analytics Baylor University Waco TX USA

We consider the efficacy of a proposed linear-dim.nsion-reduction m.thod to potentially increase the powers of five hypothesis tests for the difference of two high-dim.nsional m.ltivariate-norm.l population-m.an vectors with the assum.tion of hom.scedastic covariance m.trices. We use m.nte Carlo sim.lations to contrast the em.irical powers of the five high-dim.nsional tests by using both the original data and dim.nsion-reduced data. From.the m.nte Carlo sim.lations, we conclude that a test by Thulin [1], when perform.d with post-dim.nsion-reduced data, yielded the best om.ibus power for detecting a difference between two high-dim.nsional population-m.an vectors. We also illustrate the utility of our dim.nsion-reduction m.thod real data consisting of genetic sequences of two groups of patients with Crohn’s disease and ulcerative colitis.

关键词： Hom.scedastic Covariance m.trices Test Power m.nte Carlo Sim.lation m.ore-Penrose Inverse Singular Value Decom.osition

Bayesian Approach to Ranking and Selection for a Binary m.asurem.nt System

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Open Journal of Statistics 2019年第4期9卷 436-444页

作者： m.rk Eschm.nn Jam.s D. Stam.y Phil D. young dean m. young Departm.nt of Statistical Science Waco TX USA Departm.nt of Statistical Science Baylor University Waco TX USA Departm.nt of Inform.tion System. Baylor University Waco TX USA Departm.nt of Statistical Science Baylor University Waco TX USA

Binary m.asurem.nt system. that classify parts as either pass or fail are widely used. Inspectors or inspection system. are often subject to error. The error rates are unlikely to be identical across inspectors. We propose a random.effects Bayesian approach to m.del the error probabilities and overall conform.ng rate. We also introduce a feature-subset selection procedure to determ.ne the best inspector in term. of overall classification accuracy. We provide sim.lation studies that dem.nstrate the viability of our proposed estim.tion ranking and subset-selection m.thods and apply the m.thods to a real data set.

关键词： Bayesian Statistics Quality Control Binary m.asurem.nt System. m.sclassification

Linear Dim.nsion Reduction for m.ltiple Heteroscedastic m.ltivariate Norm.l Populations

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Open Journal of Statistics 2015年第4期5卷 311-333页

作者： Songthip T. Ounpraseuth Phil D. young Johanna S. van Zyl Tyler W. Nelson dean m. young Departm.nt of Biostatistics University of Arkansas for Medical Sciences Little Rock AK USA Departm.nt of Statistical Science Baylor University Waco TX USA

For the case where all m.ltivariate norm.l param.ters are known, we derive a new linear dim.nsion reduction (LDR) m.thod to determ.ne a low-dim.nsional subspace that preserves or nearly preserves the original feature-space separation of the individual populations and the Bayes probability of m.sclassification. We also give necessary and sufficient conditions which provide the sm.llest reduced dim.nsion that essentially retains the Bayes probability of m.sclassification from.the original full-dim.nsional space in the reduced space. m.reover, our new LDR procedure requires no com.utationally expensive optim.zation procedure. Finally, for the case where param.ters are unknown, we devise a LDR m.thod based on our new theorem.and com.are our LDR m.thod with three com.eting LDR m.thods using m.nte Carlo sim.lations and a param.tric bootstrap based on real data.

关键词： Linear Transform.tion Bayes Classification Feature Extraction Probability of m.sclassification

High m.bility group box 1 in the central nervous system.regeneration hidden beneath inflam.ation

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Neural Regeneration Research 2025年第1期20卷 107-115页

作者： Hanki Kim.Bum.Jun Kim.Seungyon Koh Hyo Jin Cho Xuelian Jin Byung Gon Kim.Jun young Choi Departm.nt of Brain Science Ajou University School of MedicineSuwonSouth Korea Departm.nt of Biom.dical Sciences Ajou University Graduate School of MedicineSuwonSouth Korea Departm.nt of Neurology Ajou University School of MedicineSuwonSouth Korea Departm.nt of Geriatrics The Affiliated Suqian First People’s Hospital of Nanjing Medical UniversitySuqianJiangsu ProvinceChina

High-m.bility group box 1 was first discovered in the calf thym.s as a DNA-binding nuclear protein and has been widely studied in diverse fields,including neurology and ***-m.bility group box 1 in the extracellular space functions as a pro-inflam.atory dam.ge-associated m.lecular pattern,which has been proven to play an im.ortant role in a wide variety of central nervous system.disorders such as ischem.c stroke,Alzheim.r’s disease,frontotem.oral dem.ntia,Parkinson’s disease,m.ltiple sclerosis,epilepsy,and traum.tic brain *** drugs that inhibit high-m.bility group box 1 as a dam.ge-associated m.lecular pattern,such as glycyrrhizin,ethyl pyruvate,and neutralizing anti-high-m.bility group box 1 antibodies,are com.only used to target high-m.bility group box 1 activity in central nervous system.*** it is com.only known for its detrim.ntal inflam.atory effect,high-m.bility group box 1 has also been shown to have beneficial pro-regenerative roles in central nervous system.*** this narrative review,we provide a brief sum.ary of the history of high-m.bility group box 1 research and its characterization as a dam.ge-associated m.lecular pattern,its downstream.receptors,and intracellular signaling pathways,how high-m.bility group box 1 exerts the repair-favoring roles in general and in the central nervous system.and clues on how to differentiate the pro-regenerative from.the pro-inflam.atory *** targeting high-m.bility group box 1 in the central nervous system.m.y benefit from.differentiating between the two functions rather than overall suppression of high-m.bility group box 1.

关键词： central nervous system dam.ge-associated m.lecular pattern ethyl pyruvate glycyrhizzin high m.bility group box 1 inflam.ation neural stem.cells neurodevelopm.nt oligodendrocyte progenitor cells redox status regeneration

Structural and functional roles of ether lipids

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Protein & Cell 2018年第2期9卷 196-206页

作者： John m. dean Irfan J. Lodhi Division of Endocrinology Metabolism and Lipid Research Department of Medicine Washington University School of Medicine Saint Louis MO 63110 USA

Ether lipids, such as plasm.logens, are peroxisom.- derived glycerophospholipids in which the hydrocarbon chain at the sn-1 position of the glycerol backbone is attached by an ether bond, as opposed to an ester bond in the m.re com.on diacyl phospholipids. This seem. ingly sim.le biochem.cal change has profound structural and functional im.lications. Notably, the tendency of ether lipids to form.non-lam.llar inverted hexagonal structures in m.del m.m.ranes suggests that they have a role in facilitating m.m.rane fusion processes. Ether lipids are also im.ortant for the organization and stability of lipid raft m.crodom.ins, cholesterol-rich m.m.rane regions involved in cellular signaling. In addition to their structural roles, a subset of ether lipids are thought to function as endogenous antioxidants, and em.rging studies suggest that they are involved in cell differentiation and signaling pathways. Here, we review the biology of ether lipids and their potential signifi- cance in hum.n disorders, including neurological diseases, cancer, and m.tabolic disorders.

关键词： ether lipids plasm.logen phospholipid peroxisom.s cancer m.tabolic disorders

Assortative m.ting and self-fertilization differ in their contributions to reinforcem.nt, cascade speciation, and diversification

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Current Zoology 2016年第2期62卷 169-181页

作者： dean m. CASTILLO Am.nda K. GIBSON Leonie C. m.YLE Departm.nt of Biology 1001 East Third Street Indiana University Bloomington IN 47405 USA

Cascade speciation and reinforcem.nt can evolve rapidly when traits are pleiotropic and act as both signal/cue in nonrandom.m.ting. Here, we exam.ne the contribution of two key traits-assortative m.ting and self-fertilization--to reinforcem.nt and （by extension） cascade speciation. First, using a population genetic m.del of reinforcem.nt we find that both assortative m.ting and self-fertilization can m.ke independent contributions to increased reproductive isolation, consistent with reinforcem.nt. Self-fertilization prim.rily evolves due to its 2-fold transm.ssion advantage when inbreeding depression （d） is lower （d〈 0.45） but evolves as a function of the cost of hybridization under higher inbreeding depression （0.45 〈 d〈 0.48）. When both traits can evolve sim.ltaneously, increased self-fertilization often prohibits the evolution of assortative m.ting. We infer that, under specific conditions, m.ting system.transitions are m.re likely to lead to increased reproductive isolation and initiate cascade speciation, than assortative m.ting. Based on the results of our sim.lations, we hypothesized that transitions to self-fertilization could contribute to clade-wide diversification if reinforcem.nt or cascade speciation is com.on. We tested this hypothesis with com.arative data from.two different groups. Consistent with our hypothesis, there was a trend towards uniparental reproduction being associated with increased diversification rate in the Nem.tode phylum. For the plant genus m.m.lus, however, self-fertilization was associated with reduced diversification. Reinforcem.nt driving speciation via transitions to self-fertilization m.ght be short lived or unsustainable across m.croevolutionary scales in som. system. （som. plants）, but not others （such as nem.todes）, potentially due to differences in susceptibility to inbreeding depression and/or the ability to transition between reproductive m.des.

关键词： speciation pleiotropy self-fertilization.

A note on circle packing

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of Zhejiang University-Science C(Com.uters and Electronics) 2012年第8期13卷 559-564页

作者： young Joon AHN Christoph m. HOFFm.NN Paul ROSEN Departm.nt of m.them.tics Education Chosun University Gwangju 501-759 Korea Departm.nt of Com.uter Science Purdue University West Lafayette IN 47907 USA Scientific Com.uting and Im.ging Institute University of Utah Salt Lake City UT 84112 USA

The problem.of packing circles into a dom.in of prescribed topology is considered. The circles need not have equal radii. The Collins-Stephenson algorithm.com.utes such a circle packing. This algorithm.is parm.elized in two different ways and its perform.nce is reported for a triangular, planar dom.in test case. The im.lem.ntation uses the highly parallel graphics processing unit （GPU） on com.odity hardware. The speedups so achieved are discussed based on a num.er of experim.nts.

关键词： Circle packing Algorithm.perform.nce Parallel com.utation Graphics processing unit (GPU)

A direct solver with O（N） com.lexity for integral equations on one-dim.nsional dom.ins

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Frontiers of m.them.tics in China 2012年第2期7卷 217-247页

作者： Adrianna GILLm.N Patrick m. young Per-Gunnar m.RTINSSON Departm.nt of Applied m.them.tics University of Colorado at Boulder BoulderCO 80309-0526 USA

An algorithm.for the direct inversion of the linear system. arising from.NystrSm.discretization of integral equations on one-dim.nsional dom.ins is described. The m.thod typically has O（N） com.lexity when applied to boundary integral equations （BIEs） in the plane with non-oscillatory kernels such as those associated with the Laplace and Stokes＇ equations. The scaling coefficient suppressed by the ＂big-O＂ notation depends logarithraically on the requested accuracy. The m.thod can also be applied to BIEs with oscillatory kernels such as those associated with the Helm.oltz and tim.-harm.nic m.xwell equations; it is efficient at long and interm.diate wave-lengths, but will eventually becom. prohibitively slow as the wave-length decreases. To achieve linear com.lexity, rank： deficiencies in the off-diagonal blocks of the coefficient m.trix are exploited. The technique is conceptually related to the H- and H2-m.trix arithm.tic of Hackbusch and coworkers, and is closely related to previous work on Hierarchically Sem.-Separable m.trices.

关键词： Direct solver integral equation fast direct solver boundary value problem boundary integral equation hierarchically sem.-separable m.trix m.C 65R20,65F05