| 《数据降维和聚类中的若干问题研究(英文版)》 1 introduction 1.1 pca and latent variable models 1.1.1 pca. 1.1.2 latent variable models 1.1.3 fa and ppca 1.2 motivations and contributions 1.3 organization of the book 2 ml estimation for factor analysis: em or non-em 2.1 introduction 2.2 fa model and three estimation algorithms 2.2.1 fa model 2.2.2 lawley (1940)'s simple iteration algorithm 2.2.3 em type algorithms 2.3 the ecme2 algorithm 2.3.1 the maximization in the first cm-step 2.3.2 the maximization in the second cm-step 2.3.3 practical consideration 2.3.4 ecme2 vs. simple iteration algorithm 2.4 the cm algorithm .2.4.1 the maximization in the second cm-step 2.4.2 when will condition i be satisfied 2.4.3 recursive computation of the matrix bi 2.4.4 on the nature of stationary points 2.5 simulations 2.5.1 simulation data 2.5.2 performance analysis 2.5.3 on different starting values 2.6 conclusion and future work 2.7 appendix 2.7.1 proofs 2.7.2 some notes 3 fast ml estimation for the mixture of factor analyzers via an ecm algorithm 3.1 introduction 3.2 mfa model and an ecm algorithm 3.2.1 the mfa model 3.2.2 the em algorithm 3.2.3 the aecm algorithm 3.2.4 the ecm algorithm 3.2.5 computational complexity 3.2.6 on speed of convergence 3.3 experiments 3.3.1 artificial data 3.3.2 real data 3.4 concluding remarks 4 mixture model selection: bic or hierarchical bic 4.1 introduction 4.2 bic and h-bic 4.2.1 bic 4.2.2 h-bic 4.3 h-bic: a large sample limit of variational bayesian lower bound 4.4 experiments 4.4.1 artificial data 4.4.2 wisconsin diagnostic breast cancer data 4.5 conclusion and discussions 4.6 appendix 4.6.1 proof of proposition 4.1 4.6.2 updating equations of mapgmm 5 a note on variational bayesian factor analysis 5.1 introduction 5.2 fa model 5.3 vbfa1 5.3.1 problem 1 of vbfa1 5.3.2 problem 2 of vbfa1 5.4 vbfa2 5.4.1 the large sample limit of vbfa2 5.4.2 backward learning of vbfa2 5.5 simulations 5.5.1 vbfa2 vs. vbfa1 5.5.2 model selection: vb vs. bic 5.6 concluding remarks and future work 5.7 appendix 5.7.1 the lower bound of vbfa2 5.7.2 justification of using constraint (5.24) on a 5.7.3 on the rotation problem of vbfa-ard 5.7.4 the large sample limit for vbfa1 6 bilinear probabilistic principal component analysis 6.1 introduction 6.1.1 motivations 6.1.2 related works 6.2 review of ppca, glram and 2dsvd 6.2.1 ppca 6.2.2 glram 6.2.3 2dsvd 6.3 bppca model 6.3.1 matrix-variate normal distribution 6.3.2 bppca model 6.4 maximum likelihood estimation of bppca 6.4.1 acm algorithm 6.4.2 an aecm algorithm 6.4.3 compression and reconstruction 6.5 connections with ppca, glram and one-mode 2dpca 6.5.1 connection between bppca and ppca 6.5.2 connection between bppca and glram 6.5.3 connection between bppca and 2dsvd 6.5.4 connection with one-mode 2dpca 6.6 experiments 6.6.1 artificial data 6.6.2 real data 6.7 conclusion 6.8 appendix 6.8.1 optimal least squares reconstruction under bppca 6.8.2 computational complexity analysis 7 conclusions and discussions references |
内容加载中，请稍后...
商品评论（0条）