DNA microarray is a powerful technology that can simultaneously determine the levels of thousands of transcripts (generated, for example, from genes/miRNAs) across different experimental conditions or tissue samples. ...
详细信息
DNA microarray is a powerful technology that can simultaneously determine the levels of thousands of transcripts (generated, for example, from genes/miRNAs) across different experimental conditions or tissue samples. The motto of differential expression analysis is to identify the transcripts whose expressions change significantly across different types of samples or experimental conditions. A number of statistical testing methods are available for this purpose. In this paper, we provide a comprehensive survey on different parametric and non-parametric testing methodologies for identifying differential expression from microarray data sets. The performances of the different testing methods have been compared based on some real-life miRNA and mRNA expression data sets. For validating the resulting differentially expressed miRNAs, the outcomes of each test are checked with the information available for miRNA in the standard miRNA database PhenomiR 2.0. Subsequently, we have prepared different simulated data sets of different sample sizes (from 10 to 100 per group/population) and thereafter the power of each test have been calculated individually. The comparative simulated study might lead to formulate robust and comprehensive judgements about the performance of each test in the basis of assumption of data distribution. Finally, a list of advantages and limitations of the different statistical tests has been provided, along with indications of some areas where further studies are required.
A major teleconnection, Atlantic multidecadal oscillation (AMO) under two phases (cool and warm) influencing precipitation extremes in Florida, USA, is the main focus of this study. Long-term extreme precipitation dat...
详细信息
A major teleconnection, Atlantic multidecadal oscillation (AMO) under two phases (cool and warm) influencing precipitation extremes in Florida, USA, is the main focus of this study. Long-term extreme precipitation data from several rain gages from temporal windows that coincide with the AMO phases are evaluated for changes in spatial and temporal variability across the region. Assessments of precipitation extremes for nine durations in different meteorologically homogenous rainfall areas as well as in the entire region are carried out. Methods of assessment included parametric unpaired t-tests and nonparametric Mann-Whitney U tests, kernel density estimates using Gaussian kernel for distribution-free comparative analysis and bootstrap sampling-based confidence intervals. Depth-duration-frequency (DDF) curves are also developed using generalized extreme value (GEV) distributions characterizing the extremes. Analysis of data indicated increase in precipitation extremes in warm phases of AMO for durations greater than 24 h. The influence of warm or cool phases of AMO on precipitation extremes is not spatially uniform in the region. Temporal shifts in occurrences of extremes from the later part of the year in warm phase to earlier in the year for the cool phase are evident. These shifts will have implications on flooding events in different regions of Florida. Magnitudes of extremes for a 25 year return period based on DDF curves were higher for all nine durations when data from cool or warm phase alone were compared to those obtained from data from two phases. Precipitation extremes for durations longer than a day are associated with increased landfalls of hurricanes occurring in the region in the AMO warm phases. (C) 2013 Elsevier B.V. All rights reserved.
This thesis analyses, derives and evaluatesspecification tests of Generalized Auto-Regressive ConditionalHeteroskedasticity (GARCH) regression models, both univariate andmultivariate. Of particular interest, in the fi...
详细信息
This thesis analyses, derives and evaluatesspecification tests of Generalized Auto-Regressive ConditionalHeteroskedasticity (GARCH) regression models, both univariate andmultivariate. Of particular interest, in the first half of thethesis, is the derivation of robust test procedures designed toassess the Constant Conditional Correlation (CCC) assumption oftenemployed in multivariate GARCH (MGARCH) models. New asymptoticallyvalid conditional moment tests are proposed which are simple toconstruct, easily implementable following the full or partial QuasiMaximum Likelihood (QML) estimation and which are robust tonon-normality. In doing so, a non-normality robust version of theTse''s (2000) LM test is provided. In addition, a new and easilyprogrammable expressions of the expected Hessian matrix associatedwith the QMLE is obtained. The finite sample performances of thesetests are investigated in an extensive Monte Carlo study,programmed in *** the second half of the thesis, attention isdevoted to nonparametric testing of GARCH regression models. Firstsimultaneous consistent nonparametrictests of the conditional meanand conditional variance structure of univariate GARCH models areconsidered. The approach is developed from the IntegratedGeneralized Spectral (IGS) and Projected Integrated ConditionalMoment (PICM) procedures proposed recently by Escanciano (2008 and2009, respectively) for time series models. Extending Escanciano(2008), a new and simple wild bootstrap procedure is proposed toimplement these tests. A Monte Carlo study compares the performanceof these nonparametrictests and four parametrictests ofnonlinearity and/or asymmetry under a wide range of *** the proposed bootstrap scheme does not strictly satisfythe asymptotic requirements, the simulation results demonstrate itsability to control the size extremely well and therefore the powercomparison seems justified. Furthermore, this suggests there mayexist weaker conditions under which th
The role of statistics in medical research starts at the planning stage of a clinical trial or laboratory experiment to establish the design and size of an experiment that will ensure a good prospect of detecting effe...
详细信息
The role of statistics in medical research starts at the planning stage of a clinical trial or laboratory experiment to establish the design and size of an experiment that will ensure a good prospect of detecting effects of clinical or scientific interest. Statistics is again used during data analysis (sample data) to make inferences valid in a wider population. In simple situations, computation of simple quantities such as P-values, confidence intervals, standard deviations, standard errors or application of some standard parametric or nonparametrictests may suffice. Moreover, despite the wide use of statistics in medical research, simple notions are sometimes misunderstood or misinterpreted by medical research workers, who have only a limited knowledge of statistics. This article, written for non-statisticians, is to explain what are the most common statistical tests used today in the field of medical research, tracing the evolution of statistical tests over time, in particular the introduction of nonparametric methods and, more recently, the nonparametric Combination (NPC) methodology. At the same time, this work seeks to identify some of the errors associated with their use, that often lead to an incorrect assessment and interpretation of results of medical research.
In this paper we study some methods to detect biased samples and to test what is the bias. These methods can be also used to obtain parametrictests for the original model. We pay special attention to the case of leng...
详细信息
In this paper we study some methods to detect biased samples and to test what is the bias. These methods can be also used to obtain parametrictests for the original model. We pay special attention to the case of length (size) biased samples. We apply these results to several real and simulated samples.
暂无评论