摘要:

This article describes a new methodology for the detection of influential subsets in regression. The method is based on an adaptation of computational and graphical techniques used in cluster analysis and makes use of some general properties of influential subsets, but it is independent of any specific measure of influence. For small to moderate data sets the proposed method is computationally efficient, compared to existing search methods, and it identifies subset candidates that merit attention according to some or all measures of joint influence that have appeared in the literature to date. Examples are given illustrating the method applied to two data sets previously analyzed in published studies.

ISSN:0040-1706    

DOI:10.1080/00401706.1984.10487980

出版商:Taylor & Francis Group    

年代:1984

数据来源: Taylor

ISSN:0040-1706    

DOI:10.1080/00401706.1984.10487981

出版商:Taylor & Francis Group    

年代:1984

数据来源: Taylor

ISSN:0040-1706    

DOI:10.1080/00401706.1984.10487982

出版商:Taylor & Francis Group    

年代:1984

数据来源: Taylor

ISSN:0040-1706    

DOI:10.1080/00401706.1984.10487983

出版商:Taylor & Francis Group    

年代:1984

数据来源: Taylor

ISSN:0040-1706    

DOI:10.1080/00401706.1984.10487984

出版商:Taylor & Francis Group    

年代:1984

数据来源: Taylor

摘要:

The best way to validate the predictive ability of a statistical model is to apply it to new data. This article compares eight ways to form regression models by forming them with old data and then validating them with fresh data. One goal here is to study which methods will work as a function of the type of data. To some extent one can tell which methods will work well by looking at the data. Another goal is to study the quality of prediction when the regression is applied to new data. Prediction quality is determined in large part by the distance of the new data in relation to the old.

ISSN:0040-1706    

DOI:10.1080/00401706.1984.10487985

出版商:Taylor & Francis Group    

年代:1984

数据来源: Taylor

摘要:

The ridge regression estimator may be written as a linear combination of the least squares estimators derived from all possible subset regressions. This article delineates the relationship between the ridge estimator and the subset regression estimators and highlights the implications of this relationship for ridge trace interpretation. Ridge-regression examples are provided, illustrating how the interpretation of a ridge trace is enhanced.

ISSN:0040-1706    

DOI:10.1080/00401706.1984.10487986

出版商:Taylor & Francis Group    

年代:1984

数据来源: Taylor

摘要:

The analysis of residuals in the calibration or inverse-regression problem is considered. A statistic is proposed that is appropriate to the specific use of the regression equation in this context—namely, estimatingxfromy. The statistic, similar in spirit to Cook's distance, may be used to measure the influence a particular observation may have on future estimates ofxfrom the calibration curve. It may also be used to measure the differences between estimates ofxobtained using calibration curves based on different models. An example is given.

ISSN:0040-1706    

DOI:10.1080/00401706.1984.10487987

出版商:Taylor & Francis Group    

年代:1984

数据来源: Taylor

摘要:

Multivariate observations obtained from chemical and biological experiments often relate several responses to a common set of parameters. In several fields of application, the response functions can be put in the form of a general multivariate linear model, which originates from the analysis of growth curves. System overdetermination is frequently encountered with this model because the number of variables that can be directly measured is larger than the number needed for full identification of the system. These measured variables usually satisfy certain equality constraints, which may come from material balances, for example. The covariate adjustment technique is particularly useful in the analysis of multiresponse models when the responses are related by equality constraints and the covariates are associated with these constraints. Methods to select covariates and criteria to evaluate the results are presented and applied to a set of experimental data. The technique yields maximum likelihood estimates of the desired parameters. The results show that the methods presented yield good parameter estimates.

ISSN:0040-1706    

DOI:10.1080/00401706.1984.10487988

出版商:Taylor & Francis Group    

年代:1984

数据来源: Taylor

摘要:

A-optimal designs for comparingvtest treatments with a control inbblocks of sizekeach are considered. Several series ofA-optimal designs are given when the parameters are in the range 2 ≤k≤ 8,k≤v≤ 30,v≤b≤ 50.A-optimal designs in blocks of size 2 are extensively studied through a combination of theoretical results and numerical investigations. Tables of approximatelyA-optimal designs are given whenA-optimal designs are not easily available for the casek= 2.

ISSN:0040-1706    

DOI:10.1080/00401706.1984.10487989

出版商:Taylor & Francis Group    

年代:1984

数据来源: Taylor