Search

In psychiatry clinical trials, an instrument or questionnaire with rating scale is often used to access safety and efficacy of a test treatment under investigation. Statistical analysis based on the collected rating scale is then performed a determine whether there is an improvement in endpoint change from baseline mean scale. This approach needs on absolute change, however, may not actually reflect the performance of the test treatment under study because the change, which may be considered of clinically importance, may fall within the same category in terms of disease severity such as mild, moderate, and severe. In this case, it is suggested, in addition to absolute change approach, a categorical shift analysis be considered to determine whether the endpoint change from baseline has resulted in an improvement in categorical shift, in terms of disease severity shift e.g., from severe to mild or moderate. In this article, we explore the relationship between absolute change approach and categorical shift analysis based on rating scales for assessment of test treatment under study. [ABSTRACT FROM AUTHOR]

We propose a modification of the pure random search algorithm for cases when the global optimum point can be located near the boundary of a feasible region. If the feasible region is cube-shaped, the worst case occurs when the global optimum point is located at the vertex of a cube. In these cases, the sample size for the pure random search should be increased significantly because the success probability of the single trial is decreased. The proposed modification is based on the virtual extension of the feasible region. The extended region contains ε -neighbourhoods of all points of the original region. All random points that fall outside the original feasible region are mapped to the nearest points on its boundary before the values of an objective function at these points are calculated. This extension of the feasible region also leads to an increase in the sample size due to the increased volume. We compare sample sizes required to hit into the neighbourhood of the global optimum point with a given probability for the proposed method and the original pure random search algorithm with a corrected sample size. We consider ball-shaped and cube-shaped feasible regions and find the sufficient conditions for the proposed algorithm to be more efficient. We offer a practical recommendation for cases, in which the feasible region is defined by a system of linear inequalities. [ABSTRACT FROM AUTHOR]

This paper investigates the problem of cost reduction of data collection procedures. To select an adequate regression or classification model, a sample set of minimum sufficient size must be collected. This sample set is modelled according to follow the data generation hypotheses. Namely, the generalised linear regression models assume the independent and identically distributed target variable. The paper analyses several numerical methods of sample size estimation and compared them in practical terms. It includes statistic, heuristic and Bayesian methods. The practical goal of a sample set collection is modelling, some methods involve analysis of the model parameters. The computational experiment includes widely-used sample sets. The open-source code and the software are provided for the practitioners to use in the data collection planning. [ABSTRACT FROM AUTHOR]

Abstract Automated cell type identification is a key computational challenge in single‐cell RNA‐sequencing (scRNA‐seq) data. To capitalise on the large collection of well‐annotated scRNA‐seq datasets, we developed scClassify, a multiscale classification framework based on ensemble learning and cell type hierarchies constructed from single or multiple annotated datasets as references. scClassify enables the estimation of sample size required for accurate classification of cell types in a cell type hierarchy and allows joint classification of cells when multiple references are available. We show that scClassify consistently performs better than other supervised cell type classification methods across 114 pairs of reference and testing data, representing a diverse combination of sizes, technologies and levels of complexity, and further demonstrate the unique components of scClassify through simulations and compendia of experimental datasets. Finally, we demonstrate the scalability of scClassify on large single‐cell atlases and highlight a novel application of identifying subpopulations of cells from the Tabula Muris data that were unidentified in the original publication. Together, scClassify represents state‐of‐the‐art methodology in automated cell type identification from scRNA‐seq data.

This work presents a statistically principled method for estimating the required number of instances in the experimental comparison of multiple algorithms on a given problem class of interest. This approach generalises earlier results by allowing researchers to design experiments based on the desired best, worst, mean or median-case statistical power to detect differences between algorithms larger than a certain threshold. Holm's step-down procedure is used to maintain the overall significance level controlled at desired levels, without resulting in overly conservative experiments. This paper also presents an approach for sampling each algorithm on each instance, based on optimal sample size ratios that minimise the total required number of runs subject to a desired accuracy in the estimation of paired differences. A case study investigating the effect of 21 variants of a custom-tailored Simulated Annealing for a class of scheduling problems is used to illustrate the application of the proposed methods for sample size calculations in the experimental comparison of algorithms. [ABSTRACT FROM AUTHOR]

Experimental comparisons of performance represent an important aspect of research on optimization algorithms. In this work we present a methodology for defining the required sample sizes for designing experiments with desired statistical properties for the comparison of two methods on a given problem class. The proposed approach allows the experimenter to define desired levels of accuracy for estimates of mean performance differences on individual problem instances, as well as the desired statistical power for comparing mean performances over a problem class of interest. The method calculates the required number of problem instances, and runs the algorithms on each test instance so that the accuracy of the estimated differences in performance is controlled at the predefined level. Two examples illustrate the application of the proposed method, and its ability to achieve the desired statistical properties with a methodologically sound definition of the relevant sample sizes. [ABSTRACT FROM AUTHOR]

Permutation or randomization test is a nonparametric test in which the null distribution (distribution under the null hypothesis of no relationship or no effect) of the test statistic is attained by calculating the values of the test statistic overall permutations (or by considering a large number of random permutation) of the observed dataset. The power of permutation test evaluated based on the observed dataset is called conditional power. In this paper, the conditional power of permutation tests is reviewed. The use of the conditional power function for sample size estimation is investigated. Moreover, reproducibility and generalizability probabilities are defined. The use of these probabilities for sample size adjustment is shown. Finally, an illustration example is used. [ABSTRACT FROM AUTHOR]

Estimation of appropriate sample size for prevalence surveys presents many challenges, particularly when the condition is very rare or has a tendency for geographical clustering. Sample size estimate for prevalence studies is a function of expected prevalence and precision for a given level of confidence expressed by the z statistic. Choice of the appropriate values for these variables is sometimes not straight-forward. Certain other situations do not fulfil the assumptions made in the conventional equation and present a special challenge. These situations include, but are not limited to, smaller population size in relation to sample size, sampling technique or missing data. This paper discusses practical issues in sample size estimation for prevalence studies with an objective to help clinicians and healthcare researchers make more informed decisions whether reviewing or conducting such a study. [ABSTRACT FROM AUTHOR]

To determine the reproducibility of measurements of brain 5-HT2A receptors with an [18F]altanserin PET bolus/infusion approach. Further, to estimate the sample size needed to detect regional differences between two groups and, finally, to evaluate how partial volume correction affects reproducibility and the required sample size. For assessment of the variability, six subjects were investigated with [18F]altanserin PET twice, at an interval of less than 2 weeks. The sample size required to detect a 20% difference was estimated from [18F]altanserin PET studies in 84 healthy subjects. Regions of interest were automatically delineated on co-registered MR and PET images. In cortical brain regions with a high density of 5-HT2A receptors, the outcome parameter (binding potential, BP1) showed high reproducibility, with a median difference between the two group measurements of 6% (range 5–12%), whereas in regions with a low receptor density, BP1 reproducibility was lower, with a median difference of 17% (range 11–39%). Partial volume correction reduced the variability in the sample considerably. The sample size required to detect a 20% difference in brain regions with high receptor density is approximately 27, whereas for low receptor binding regions the required sample size is substantially higher. This study demonstrates that [18F]altanserin PET with a bolus/infusion design has very low variability, particularly in larger brain regions with high 5-HT2A receptor density. Moreover, partial volume correction considerably reduces the sample size required to detect regional changes between groups. [ABSTRACT FROM AUTHOR]

Federal regulations require operators of nuclear and coal-fired power-generating stations to estimate the number of fish impinged on intake screens. During winter months, impingement may range into the hundreds of thousands for certain species, making it impossible to count all intake screens completely. We present graphs for determinig the appropriate'optimal' subsample that must be obtained to estimate the total number impinged. Since the number of fish impinged tends to change drastically within a short time period, the subsample size is determined based on the most recent data. This allows for the changing nature of the species-age composition of the impinged fish. These graphs can also be used for subsampling fish catches in an aquatic system when the size of the catch is too large to sample completely. [ABSTRACT FROM AUTHOR]