Explore the vast range of titles available.

The simultaneous estimation and variable selection for Cox model has been discussed by several authors when one observes right-censored failure time data. However, there does not seem to exist an established procedure for interval-censored data, a more general and complex type of failure time data, except two parametric procedures. To address this, we propose a broken adaptive ridge (BAR) regression procedure that combines the strengths of the quadratic regularization and the adaptive weighted bridge shrinkage. In particular, the method allows for the number of covariates to be diverging with the sample size. Under some weak regularity conditions, unlike most of the existing variable selection methods, we establish both the oracle property and the grouping effect of the proposed BAR procedure. An extensive simulation study is conducted and indicates that the proposed approach works well in practical situations and deals with the collinearity problem better than the other oracle-like methods. An application is also provided.

We propose the elastic net, a new regularization and variable selection method. Real world data and a simulation study show that the elastic net often outperforms the lasso, while enjoying a similar sparsity of representation. In addition, the elastic net encourages a grouping effect, where strongly correlated predictors tend to be in or out of the model together. The elastic net is particularly useful when the number of predictors (p) is much bigger than the number of observations (n). By contrast, the lasso is not a very satisfactory variable selection method in the p ≫ n case. An algorithm called LARS-EN is proposed for computing elastic net regularization paths efficiently, much like algorithm LARS does for the lasso.

Multi-view subspace clustering has attracted extensive attention in recent years due to the fact that it can utilize the self-expressive property to reveal the low-dimensional subspace segmentation. However, due to the noise corruptions in the real-word applications, the learned affinity matrix may not be able to reflect the intrinsic correlations among samples faithfully. Moreover, most self-expressive methods only focus on exploring the global structure of each view, but overlook the local structure information, especially the complex manifold structure. To overcome these problems, we propose a novel algorithm termed L ocality-constrained D ouble-layer S tructure S caled S implex (LDSSS) multi-view subspace clustering, which can simultaneously capture the global and local structure of each view. For the global structure, a double-layer self-expressive structure is introduced to realize relation reconstruction, which can reveal the credible correlations and mitigate the noise effect. The view-specific grouping effect based on the graph Laplacian is adopted to preserve the local structure of data in each view. Moreover, with regard to the more complex manifold structure, we also discuss another two types of manifold regularizations, Hessian regularization and hypergraph Laplacian, conducting the view-specific grouping effect respectively. The scaled simplex representation and the consensus representation are integrated into a joint framework to balance the scale of different views and directly obtain a non-negative and sparse consensus matrix. Finally, an efficient optimization algorithm based on the alternating direction method of multipliers (ADMM) is designed to solve LDSSS. Experimental results on eight datasets demonstrate that the proposed method has better clustering performance than other multi-view subspace clustering methods. The source code is released on Github: https://github.com/asdhff/LDSSS .

Broken adaptive ridge (BAR) is a computationally scalable surrogate to L 0 -penalized regression, which involves iteratively performing reweighted L 2 penalized regressions and enjoys some appealing properties of both L 0 and L 2 penalized regressions while avoiding some of their limitations. In this paper, we extend the BAR method to the semi-parametric accelerated failure time (AFT) model for right-censored survival data. Specifically, we propose a censored BAR (CBAR) estimator by applying the BAR algorithm to the Leurgan’s synthetic data and show that the resulting CBAR estimator is consistent for variable selection, possesses an oracle property for parameter estimation and enjoys a grouping property for highly correlation covariates. Both low- and high-dimensional covariates are considered. The effectiveness of our method is demonstrated and compared with some popular penalization methods using simulations. Real data illustrations are provided on a diffuse large-B-cell lymphoma data and a glioblastoma multiforme data.

The standard L2-norm support vector machine (SVM) is a widely used tool for classification problems. The L1-norm SVM is a variant of the standard L2-norm SVM, that constrains the L1-norm of the fitted coefficients. Due to the nature of the L1-norm, the L1-norm SVM has the property of automatically selecting variables, not shared by the standard L2-norm SVM. It has been argued that the L1-norm SVM may have some advantage over the L2-norm SVM, especially with high dimensional problems and when there are redundant noise variables. On the other hand, the L1-norm SVM has two drawbacks: (1) when there are several highly correlated variables, the L1-norm SVM tends to pick only a few of them, and remove the rest; (2) the number of selected variables is upper bounded by the size of the training data. A typical example where these occur is in gene microarray analysis. In this paper, we propose a doubly regularized support vector machine (DrSVM). The DrSVM uses the elastic-net penalty, a mixture of the L2-norm and the L1-norm penalties. By doing so, the DrSVM performs automatic variable selection in a way similar to the L1-norm SVM. In addition, the DrSVM encourages highly correlated variables to be selected (or removed) together. We illustrate how the DrSVM can be particularly useful when the number of variables is much larger than the size of the training data (p ⨠ n). We also develop efficient algorithms to compute the whole solution paths of the DrSVM.

While a testing sample may be first encoded linearly with labeled samples and then classified with KNN on the sum of the obtained weights of the samples in each class so as to avoid the consistent distribution assumption explicitly or implicitly used in most of the existing classification methods for training and testing samples, a novel self-paced and Bayes-decision-rule linear KNN prediction method SBLD-KNN in this study will be proposed to achieve three goals: (1) class-ware information will be explicitly reflected in a grouping effect regularization term so as to share the sparsity of a linear encoder and simultaneously have grouping effect of weights on each class; (2) the resultant predictor behaves like Bayes-decision-rule for minimum error; (3) self-paced regularized term is designed to adaptively truncate the weights of labeled samples for enhancing generalization. In order to do so, the corresponding objective function of SBLD-KNN is designed and then optimized by using the alternating optimization strategy, and its Bayes-decision-rule is theoretically analyzed. Our experimental results on benchmark datasets witness the effectiveness of SBLD-KNN, in contrast to the comparative methods, including SBLD-KNN’s simplified version BD-KNN with weight’s truncating rather than self-pacing.

Least squares regression (LSR) is an effective method that has been widely used for subspace clustering. Under the conditions of independent subspaces and noise-free data, coefficient matrices can satisfy enforced block diagonal (EBD) structures and achieve good clustering results. More importantly, LSR produces closed solutions that are easier to solve. However, solutions with block diagonal properties that have been solved using LSR are sensitive to noise or corruption as they are fragile and easily destroyed. Moreover, when using actual datasets, these structures cannot always guarantee satisfactory clustering results. Considering that block diagonal representation has excellent clustering performance, the idea of block diagonal constraints has been introduced into LSR and a new subspace clustering method, which is named block diagonal least squares regression (BDLSR), has been proposed. By using a block diagonal regularizer, BDLSR can effectively reinforce the fragile block diagonal structures of the obtained matrices and improve the clustering performance. Our experiments using several real datasets illustrated that BDLSR produced a higher clustering performance compared to other algorithms.

This paper is devoted to identifying the biomarkers of rat liver regeneration via the adaptive logistic regression. By combining the adaptive elastic net penalty with the logistic regression loss, the adaptive logistic regression is proposed to adaptively identify the important genes in groups. Furthermore, by improving the pathwise coordinate descent algorithm, a fast solving algorithm is developed for computing the regularized paths of the adaptive logistic regression. The results from the experiments performed on the microarray data of rat liver regeneration are provided to illustrate the effectiveness of the proposed method and verify the biological rationality of the selected biomarkers.

Abstract Background Novel feline RBC antigens might contribute to decreased efficacy of RBC transfusion and increased incidence of acute transfusion reactions (ATR). Objectives To examine the effect of major cross-match in transfusion-naïve anemic cats on the incidence of acute immunologic transfusion reaction and transfusion efficacy for up to 24 hours after transfusion. Animals Forty-eight client owned transfusion-naïve anemic cats. Methods Prospective, randomized, controlled study. All transfusion-naïve cats receiving packed red blood cells (pRBC) transfusions from January 2016 to August 2017 were eligible for inclusion. Cats in the study group received cross-match and blood type compatible pRBCs and cats in the control group received noncross-matched blood type compatible pRBCs. Incidence of ATR and change in PCV after transfusion was recorded. Results No significant difference in incidence of transfusion reactions between cross-matched and noncross-matched groups (CM+ 4/24; 17%, CM– 7/24; 29%, P = .16). No significant difference between groups in mean change in PCV after transfusion scaled to dose of pRBCs administered at any time point after transfusion (immediate: CM+ 0.62 ± 0.59, CM– 0.75 ± 0.48, P = .41; 1 hour: CM+ 0.60 ± 0.66, CM– 0.74 ± 0.53, P = .43; 12 hours: CM+ 0.70 ± 0.55, CM– 0.66 ± 0.60, P = .81; 24 hours: CM+ 0.64 ± 0.71, CM– 0.55 ± 0.48, P = .70). Conclusions and Clinical Importance Our results do not support use of the major cross-match test to increase efficacy of, and to decrease adverse events associated with, RBC transfusion in AB blood typed transfusion-naïve cats.

Language learning requires a capacity to recall novel series of speech sounds. Research shows that prosodic marks create grouping effects enhancing serial recall. However, any restriction on memory affecting the reproduction of prosody would limit the set of patterns that could be learned and subsequently used in speech. By implication, grouping effects of prosody would also be limited to reproducible patterns. This view of the role of prosody and the contribution of memory processes in the organization of prosodic patterns is examined by evaluating the correspondence between a reported tendency to restrict stress intervals in speech and size limits on stress-grouping effects. French speech is used where stress defines the endpoints of groups. In Experiment 1, 40 speakers recalled novel series of syllables containing stress-groups of varying size. Recall was not enhanced by groupings exceeding four syllables, which corresponded to a restriction on the reproducibility of stress-groups. In Experiment 2, the subjects produced given sentences containing phrases of differing length. The results show a strong tendency to insert stress within phrases that exceed four syllables. Since prosody can arise in the recall of syntactically unstructured lists, the results offer initial support for viewing memory processes as a factor of stress-rhythm organization.