Clustering of gene expression data is widely used to identify novel subtypes cancer. Plenty clustering approaches have been proposed, but there a lack knowledge regarding their relative merits and how characteristics influence the performance. We evaluate cluster analysis choices affect performance by studying four publicly available human cancer sets: breast, brain, kidney stomach In particular, we focus on sample size, distribution heterogeneity performance.In general, increasing size had...

Cancer subtype identification is important to facilitate cancer diagnosis and select effective treatments. Clustering of patients based on high-dimensional RNA-sequencing data can be used detect novel subtypes, but only a subset the features (e.g., genes) contains information related subtype. Therefore, it reasonable assume that clustering should set carefully selected rather than all features. Several feature selection methods have been proposed, how when use these are still poorly...

Entropy-type integral functionals of densities are widely used in mathematical statistics, information theory, and computer science. Examples include measures closeness between distributions (e.g., density power divergence) uncertainty characteristics for a random variable Rényi entropy). In this paper, we study U-statistic estimators class such functionals. The based on ε-close vector observations the corresponding independent identically distributed samples. We prove asymptotic properties...

The Rényi entropy is a generalisation of the Shannon and widely used in mathematical statistics applied sciences for quantifying uncertainty probability distribution. We consider estimation quadratic related functionals marginal distribution stationary m-dependent sequence. U-statistic estimators under study are based on number ε-close vector observations corresponding sample. A variety asymptotic properties these obtained (e.g. consistency, normality, Poisson convergence). results can be...

Abstract Background Metastasized clear cell renal carcinoma (ccRCC) is associated with a poor prognosis. Almost one-third of patients non-metastatic tumors at diagnosis will later progress metastatic disease. These need to be identified already diagnosis, undertake closer follow up and/or adjuvant treatment. Today, clinicopathological variables are used risk classify patients, but molecular biomarkers needed improve classification identify the high-risk which benefit most from modern...

Weighting methods are used in observational studies to adjust for covariate imbalances between treatment and control groups. Entropy balancing (EB) is an alternative inverse probability weighting with estimated propensity score. The EB weights constructed satisfy balance constraints optimized towards stability. Large sample properties of estimators the average causal effect, based on Kullback-Leibler quadratic Rényi relative entropies, described. Additionally, their asymptotic variances...

Matching a query (reference) image to an extracted from database containing (possibly) transformed copies is important retrieval task. In this paper we present general method based on matching densities of the corresponding feature vectors by using Bregman distances. We consider statistical estimators for some quadratic entropy-type characteristics. particular, distances can be evaluated in problems whenever images are modeled random large databases. Moreover, used average case analysis...

Entropy-type integral functionals of densities are widely used in mathematical statistics, information theory, and computer science. Examples include measures closeness between distributions (e.g., density power divergence) uncertainty characteristics for a random variable R\'enyi entropy). In this paper, we study U-statistic estimators class such functionals. The based on epsilon-close vector observations the corresponding independent identically distributed samples. We prove asymptotic...

In clustering of high-dimensional data a variable selection is commonly applied to obtain an accurate grouping the samples. For two-class problems this may be carried out by fitting mixture distribution each variable. We propose hybrid method for estimating parametric two symmetric densities. The estimator combines moments with minimum distance approach. An evaluation study including both extensive simulations and gene expression from acute leukemia patients shows that outperforms...

Abstract Clustering of gene expression data is widely used to identify novel subtypes cancer. Plenty clustering approaches have been proposed, but there a lack knowledge regarding their relative merits and how characteristics influence the performance. We evaluate cluster analysis choices affect performance by studying four publicly available human cancer sets: breast, brain, kidney stomach In particular, we focus on sample size, distribution heterogeneity general, increasing size had...

In this paper, we study estimation of certain integral functionals one or two densities with samples from stationary m-dependent sequences. We consider types U-statistic estimators for these that are functions the number epsilon-close vector observations in samples. show consistent and obtain their rates convergence under weak distributional assumptions. particular, propose based on incomplete U-statistics which have favorable consistency properties even when m-dependence is only dependence...

Numerous entropy-type characteristics (functionals) generalizing R\'enyi entropy are widely used in mathematical statistics, physics, information theory, and signal processing for characterizing uncertainty probability distributions distribution identification problems. We consider estimators of some (integral) functionals discrete continuous based on the number epsilon-close vector records corresponding independent identically distributed samples from two distributions. The form a...

The R\'enyi entropy is a generalization of the Shannon and widely used in mathematical statistics applied sciences for quantifying uncertainty probability distribution. We consider estimation quadratic related functionals marginal distribution stationary m-dependent sequence. U-statistic estimators under study are based on number epsilon-close vector observations corresponding sample. A variety asymptotic properties these obtained (e.g., consistency, normality, Poisson convergence). results...

Abstract Background: RNA-seq data from tumor samples can be used to identify novel cancer subtypes using cluster analysis. The number of features is often large compared the and different clusters appear in subsets feature space. Feature selection techniques are therefore commonly reduce dimension remove redundant irrelevant before performing An abundance methods have been proposed literature, but it unclear how ability analysis affected by choice method. Method: We evaluated 13 on four...