摘要:

Multivariate control charts using Hotelling'sT2statistic are popular and easy to use but interpreting their signals can be a problem. Identifying which characteristic or group of characteristics is out of control when the chart signals often necessitates an examination of the univariate charts for each variable. It is shown in this paper that the interpretation of a signal from aT2statistic is greatly aided if the corresponding value is partitioned into independent parts. Information on which characteristic is significantly contributing to the signal is readily available from this decomposition.

ISSN:0022-4065    

DOI:10.1080/00224065.1995.11979573

出版商:Taylor&Francis    

年代:1995

数据来源: Taylor

摘要:

Cumulative sum (CUSUM) control charts have been widely used for monitoring the process mean. Relatively little attention has been given to the use of CUSUM charts for monitoring the process variance. The properties of CUSUM charts based on the logarithmic transformation of the sample variance log(S2) for detecting changes in the process variance are described in this paper. Design procedures are developed for designing one-sided and two-sided CUSUM charts that are nearly optimal. The fast initial response feature as an enhancement to the CUSUM chart is discussed. The effects of non-normality and serially correlated observations on the performance of the CUSUM chart is examined. The relative performance of CUSUM charts based on the sample varianceS2and log(S2) is also examined in this paper. A comparison of the performances of CUSUM and exponentially weighted moving average (EWMA) charts based on log(S2) is presented. It is found that the performance of a CUSUM chart is comparable to that of the corresponding EWMA chart.

ISSN:0022-4065    

DOI:10.1080/00224065.1995.11979574

出版商:Taylor&Francis    

年代:1995

数据来源: Taylor

摘要:

This paper proposes a heuristic method based on a weighted variance concept of setting up control limits ofX̄andRcharts for skewed populations. It provides asymmetric control limits in accordance with the direction and degree of skewness estimated from the sample data, by using different variances in computing upper and lower control limits. For symmetric populations, however, these control limits are equivalent to those of Shewhart control charts. The new heuristic control charts are compared by Monte Carlo simulation with Shewhart charts and the geometric control charts of Ferrell. When the underlying population is Weibull or Burr's, the heuristic charts are found to perform better than Shewhart or geometric charts as the skewness increases.

ISSN:0022-4065    

DOI:10.1080/00224065.1995.11979575

出版商:Taylor&Francis    

年代:1995

数据来源: Taylor

摘要:

The problem of testing equality of several normal means to a specifed value when the population variances are unknown is considered. Two test procedures, one obtained by using Fisher's method of combining independent tests and the other based on the maximum of independenttstatistics are proposed. In addition, the likelihood ratio test statistic and its asymptotic null distribution are derived. Power comparisons of these three test procedures are made using simulation. Assuming that the unknown variances are equal, these tests are again compared with an appropriately modified analysis of variance procedure. Based on the simulation study, the likehood ratio test and Fisher's method are recommended for use in general. When the variances are equal, the modified analysis of variance test is superior.

ISSN:0022-4065    

DOI:10.1080/00224065.1995.11979576

出版商:Taylor&Francis    

年代:1995

数据来源: Taylor

摘要:

The most commonly used techniques in statistical process control are parametric, and so they require assumptions regarding the statistical properties of the underlying process. For example, Shewhart control charts assume that the observations are independent, and that the statistic of interest is normally distributed. These assumptions are often violated in practice; for example, the distribution of the variable being measured may be strongly skewed or may fail a test for normality. In such cases the control limits, especially for small subgroup samples, may not be accurate. The bootstrap is a computer intensive resampling procedure that does not require aprioridistribution assumptions. It was developed to find the distribution of a statistic when the distribution is not known. We first extend the bootstrap percentile method to include a series of subgroups, which are typically used in assessing process control limits. We show, via examples, how the subgroup bootstrap is used to assess process control limits forX̄andS2charts. Via simulation, we then empirically compare the subgroup bootstrap and parametric methods for determining process control limits for a quality related characteristic of a manufacturing process under various conditions. The results show that bootstrap methods forX̄andS2control charts generally achieve comparatively better control limit estimates than standard parametric methods, particularly when the assumption of a normal process distribution is not valid. The subgroup bootstrap is easily implemented on a personal computer as a general methodology for statistical process control, and hence, is a potentially useful pragmatic quality improvement tool.

ISSN:0022-4065    

DOI:10.1080/00224065.1995.11979577

出版商:Taylor&Francis    

年代:1995

数据来源: Taylor

摘要:

Skewed distributions play an important role in the analysis of data from quality and reliability experiments. Very often unknown parameters must be estimated from the sample data in order to test whether the data has come from a certain family of distributions. Because a shape parameter appears in most skewed distributions, this goodness-of-fit test can be difficult to perform and may require extensive tables. In this paper, we propose a general purpose approximate goodness-of-fit test which requires very few critical points, is easy to carry out, and can be used to test for the validity of different families of skewed distributions.

ISSN:0022-4065    

DOI:10.1080/00224065.1995.11979578

出版商:Taylor&Francis    

年代:1995

数据来源: Taylor

摘要:

The exponentially weighted moving average (EWMA) of past data is frequently used in process control applications. In engineering process control, the mean level of the quality characteristic is assumed to wander over time. If the integrated moving average IMA(0,1,1) model is used to represent a process disturbance, the EWMA of the past data has optimal properties as a forecast of the next observation. Advantages of this model for representing the process disturbance have been discussed by Box and Kramer. Using the IMA(0,1,1) model, Box and Kramer have also considered a model for feedback control which takes account of the costs of making an adjustment, of taking a sample, and of being off target. Unfortunately, it is sometimes difficult to estimate, using past process data, the smoothing constant needed to update the EWMA of past data. In this article, a simple method for obtaining the maximum likelihood estimate and a confidence interval for the smoothing constant is discussed, and an accurate and efficient computer routine is provided for performing the computations. Two examples which use actual chemical process data are presented. The 95% confidence interval for the smoothing constant is markedly different in both examples, demonstrating the importance of the information contained in past process data.

ISSN:0022-4065    

DOI:10.1080/00224065.1995.11979579

出版商:Taylor&Francis    

年代:1995

数据来源: Taylor

ISSN:0022-4065    

DOI:10.1080/00224065.1995.11979580

出版商:Taylor&Francis    

年代:1995

数据来源: Taylor

ISSN:0022-4065    

DOI:10.1080/00224065.1995.11979581

出版商:Taylor&Francis    

年代:1995

数据来源: Taylor

ISSN:0022-4065    

DOI:10.1080/00224065.1995.11979582

出版商:Taylor&Francis    

年代:1995

数据来源: Taylor