摘要:

The outlying tendency of any case in a multiple regression ofppredictors may be estimated by drawing all subsets of sizepfrom the remaining cases and fitting the model. Each such subset yields anelemental residualfor the case in question, and a suitable summary statistic of them can be used as an estimate of the case's outlying tendency. We propose two such summary statistics: an unweighted median, which is of bounded influence, and a weighted median, which is more efficient but less robust. The computational load of the procedure is reduced by using random samples in place of the full set of subsets of sizep. As a byproduct the method yields useful information on the influence (or leverage) of cases and the mutual masking of high leverage points.

ISSN:0040-1706    

DOI:10.1080/00401706.1984.10487956

出版商:Taylor & Francis Group    

年代:1984

数据来源: Taylor

摘要:

Constraints, in the form of upper and lower bounds, are often placed on the components of mixtures. These constraints restrict the size of the experimental region and may alter its shape. A transformation from the original components to new variables, called pseudocomponents, is typically used to reduce the ill conditioning created by the restricted size of the experimental region. If the experimental region is a simplex, the pseudocomponent transformation is also used to adapt standard designs to the experimental region. A new pseudocomponent transformation is introduced, and the conditions under which it is preferable to the usual pseudocomponent transformation are given. The effect of changing the constraints on both the size and shape of the experimental region is discussed.

ISSN:0040-1706    

DOI:10.1080/00401706.1984.10487957

出版商:Taylor & Francis Group    

年代:1984

数据来源: Taylor

摘要:

In experiments for response estimation, algorithms developed for constructingD-optimal exact (integer replication) designs may be inappropriate when the number of observations is not large relative to the number of parameters. This article generalizes Mitchell's DETMAX algorithm to an arbitrary design criterion and describes an efficient implementation for minimizing either the maximum variance or the average variance of the response estimator over the design region. Some examples are given that demonstrate the possible improvements in response estimator variances.

ISSN:0040-1706    

DOI:10.1080/00401706.1984.10487958

出版商:Taylor & Francis Group    

年代:1984

数据来源: Taylor

摘要:

Fries and Hunter (1980) presented a practical algorithm for selecting standard 2n–mfractional factorial designs based on a criterion they called “minimum aberration.” In this article some simple results are presented that enable the Fries–Hunter algorithm to be used for a wider range ofnandmand for designs with factors atplevels wherep≥ 2 is prime. Examples of minimum aberration 2n–mdesigns with resolutionR≥ 4 are given form,n–m< 9. A matrix is given for generating 3n–mdesigns withm,n–m≤ 6, which have, or nearly have, minimum aberration.

ISSN:0040-1706    

DOI:10.1080/00401706.1984.10487959

出版商:Taylor & Francis Group    

年代:1984

数据来源: Taylor

摘要:

Methods are presented for calibrating when there are systematic departures from an exact linear model. The models proposed are much less structured; nonetheless, they admit uncertainty statements about calibration intervals that are usable. Application is made to pressure-volume calibration of a nuclear accountability tank.

ISSN:0040-1706    

DOI:10.1080/00401706.1984.10487960

出版商:Taylor & Francis Group    

年代:1984

数据来源: Taylor

摘要:

We consider the problem of summarizing a scatterplot with a smooth, monotone curve. A solution that combines local averaging and isotonic regression is proposed, and we demonstrate its use with two examples. In the second example, the procedure is applied, in a regression setting, to some data from Box and Cox (1964), and it is shown how this new procedure generalizes Box and Cox's well-known family of transformations. In the same example, the bootstrap is applied to obtain a measure of the variability of the procedure.

ISSN:0040-1706    

DOI:10.1080/00401706.1984.10487961

出版商:Taylor & Francis Group    

年代:1984

数据来源: Taylor

摘要:

This article discusses procedures for computing tolerance limits for a population consisting of many batches. The results of Mee and Owen (1983) for one-sided β-content tolerance limits are extended to the two-sided case. In addition, β-expectation tolerance limit procedures are presented. The procedures given for the case in which the ratio of thebetweenandwithinvariances is known are exact, whereas the procedures derived for settings in which the ratio is unknown involve an approximation.

ISSN:0040-1706    

DOI:10.1080/00401706.1984.10487962

出版商:Taylor & Francis Group    

年代:1984

数据来源: Taylor

摘要:

Researchers are often confronted with problems requiring consideration of a large number of factors. Because such situations require large and costly experimental programs, an efficient method of factor screening is often needed. This article examines the performance characteristics of two-stage group screening experiments. To evaluate performance, random grouping and a constant signal-to-noise ratio for all active effects are assumed. No presumption is made concerning the directions of possible effects. As part of the investigation, a computer-aided search routine is developed to provide guidance in selecting and using a satisfactory two-stage group screening strategy. Important practical considerations are also discussed.

ISSN:0040-1706    

DOI:10.1080/00401706.1984.10487963

出版商:Taylor & Francis Group    

年代:1984

数据来源: Taylor

摘要:

Consider the problem of obtaining simultaneous confidence bands for the survival function S(x) when the data are arbitrarily right censored. The usual pointwise confidence intervals based on Greenwood's variance formula can be adapted to yield a large-sample confidence band. This band has, in a certain sense, equal precision at each point of S(x). It is compared with the censored versions of the Kolmogorov band and the Rényi band. The comparisons are made in terms of the widths and the adequacy of large-sample approximations and are carried out under various censoring models and degrees of censoring. The bands are illustrated by applying them to data from a mechanical-switch life test.

ISSN:0040-1706    

DOI:10.1080/00401706.1984.10487964

出版商:Taylor & Francis Group    

年代:1984

数据来源: Taylor

摘要:

This article investigates the consequences of departures from independence when the component lifetimes in a series system are exponentially distributed. Such departures are studied when the joint distribution is assumed to follow a Gumbel bivariate exponential model. Two distinct situations are considered. First, in theoretical modeling of series systems, when the distribution of the component lifetimes is assumed, one wishes to compute system reliability and mean system life. Second, errors in parametric and nonparametric estimation of component reliability and component mean life are studied based on life-test data collected on series systems when the assumption of independence is made erroneously. Systems with two components are studied.

ISSN:0040-1706    

DOI:10.1080/00401706.1984.10487965

出版商:Taylor & Francis Group    

年代:1984

数据来源: Taylor