ISSN:0040-1706    

DOI:10.1080/00401706.1994.10485392

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

We derive run-length distributions of the special-cause control chart proposed by Alwan and Roberts for correlated observations, given that the assignable cause to be detected is a shift in the process mean. Both recursive and closed-form solutions are derived for the run-length distribution, average run length (ARL), and standard deviation of the run length (SRL) for any AR(p) process, and approximate solutions are derived for the more general ARMA(p,q) processes. The expressions derived do not depend on the type of shift in the process mean. Numerical results are illustrated for the ARL and SRL of the ARMA(l,l) model, given that the shift in the mean is a step shift. These results show that the ARL and SRL of the specialcause control chart are relatively smaller when the process is negatively rather than positively autocorrelated. Regardless of the sign of the autocorrelation, the shape of the probability mass function of the run length reveals that the probability of detecting shifts very early is substantially higher for the special-cause chart than for more traditional control charts. Early detection makes the cause of the signal easier to identify, resulting in a more rapid rate of continuous quality improvement. There are some cases, however, when traditional charts, which are simpler to implement, should be considered even when the process is autocorrelated.

ISSN:0040-1706    

DOI:10.1080/00401706.1994.10485393

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

ISSN:0040-1706    

DOI:10.1080/00401706.1994.10485394

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

ISSN:0040-1706    

DOI:10.1080/00401706.1994.10485395

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

ISSN:0040-1706    

DOI:10.1080/00401706.1994.10485396

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

ISSN:0040-1706    

DOI:10.1080/00401706.1994.10485397

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

The problem of analyzing weight frequency particle-size data is considered with respect to estimation of the sampling error of the size fractions by weight. The established nonparametric method of Gy is reassessed in terms of a general framework that allows simultaneous consideration of both parametric and nonparametric approaches to the problem. Two additional advantages result from introducing the present general framework. First, the formulas for the sampling variance of the sample size fraction by weight can be readily deduced from a genera1 approach due to Scheaffer. Second, a fully parameteric approach can be developed for the case when a lognormal parametric mode1 is appropriate for the weight frequency distribution of the particles. For other parametric models, an approximate parametric approach is indicated. For the nonparametric case, an approach, alternative to Gy's, is also obtained. In all approaches developed in this article, replication data on the observed weightfrequency particle-size distribution is assumed to be available.

ISSN:0040-1706    

DOI:10.1080/00401706.1994.10485398

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

D-optimal and other computer-generated experimental designs have been criticized for being too dependent on an assumed statistical model. To address this criticism, we introduce the notion of empirical models that have both primary and potential terms. Combining this idea with the Bayesian paradigm, this article proposes a modification of the D-optimal approach that preserves the flexibility and ease of use of algorithmic designs while being more resistant to the biases caused by an incorrect model. These designs provide a Bayesian justification for resolution IV designs. Several theoretical examples and a practical example from the literature demonstrate the advantages of the proposed method.

ISSN:0040-1706    

DOI:10.1080/00401706.1994.10485399

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

In reliability applications, it is often of interest to estimate extreme quantiles of the distribution of critical stimulus levels based on a limited number of sensitivity tests. To evaluate and compare the effectiveness of available sequential-design sensitivity tests in this setting, several such tests were compared through Monte Carlo simulation. Samples of size 20, 35, and 50 were used to estimate the .99 and, 999 quantiles of the probit model. Each test was evaluated using initial parameter estimates equal to the parameters of the model and then using inaccurate estimates. The effect of assuming the probit model when the true model was logit, generalized logistic, or Type I extreme-value was also studied. Our results show that, when the model is correctly specified, tests designed to estimate the model parameters and subsequently to estimate quantiles as a function of the model parameters provided more accurate quantile estimates than tests designed to estimate a specified quantile. The tests that estimated model parameters, however, were more affected by model misspecification.

ISSN:0040-1706    

DOI:10.1080/00401706.1994.10485400

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

Sensitivity tests are often used to estimate the parameters associated with latent continuous variables that cannot be measured. For example, each explosive specimen has a threshold. The specimen will detonate if and only if an applied shock exceeds this value. Since there is no way to determine the threshold of an individual, specimens are tested at various levels to determine parameters of the population. A new test described here produces efficient estimates of the parameters of the distribution, even with limited prior knowledge. This test efficiently characterizes the entire distribution and desired percentiles of any population.

ISSN:0040-1706    

DOI:10.1080/00401706.1994.10485401

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor