Adaptive Regression for Modeling Nonlinear Relationships by George J. Knafl, Kai Ding

By George J. Knafl, Kai Ding

This publication provides tools for investigating even if relationships are linear or nonlinear and for adaptively becoming applicable types once they are nonlinear. facts analysts will the right way to contain nonlinearity in a single or extra predictor variables into regression versions for various different types of consequence variables. Such nonlinear dependence is frequently no longer thought of in utilized learn, but nonlinear relationships are universal and so must be addressed. a customary linear research can produce deceptive conclusions, whereas a nonlinear research gives you novel insights into information, no longer another way attainable.

A number of examples of the advantages of modeling nonlinear relationships are provided in the course of the publication. tools are lined utilizing what are referred to as fractional polynomials according to real-valued strength alterations of basic predictor variables mixed with version choice in response to probability cross-validation. The ebook covers easy methods to formulate and behavior such adaptive fractional polynomial modeling within the general, logistic, and Poisson regression contexts with non-stop, discrete, and counts results, respectively, both univariate or multivariate. The publication additionally offers a comparability of adaptive modeling to generalized additive modeling (GAM) and a number of adaptive regression splines (MARS) for univariate results.

The authors have created personalized SAS macros to be used in undertaking adaptive regression modeling. those macros and code for carrying out the analyses mentioned within the ebook can be found throughout the first author's site and on-line through the book’s Springer web site. precise descriptions of the way to exploit those macros and interpret their output seem during the booklet. those equipment may be carried out utilizing different courses.

Show description

Read or Download Adaptive Regression for Modeling Nonlinear Relationships PDF

Similar biostatistics books

Statistics for Veterinary and Animal Science, Second Edition

Reading statistical info and engaging in statistical methods is an essential component of veterinary and animal technological know-how. This profitable textbook, now in its moment version and with an accompanying CD, presents the reader with the mandatory info to deal with numerical info and significantly appraise the statistical method within the literature within the fields of veterinary and animal technology.

Frailty Models in Survival Analysis (Chapman & Hall CRC Biostatistics Series)

The concept that of frailty bargains a handy option to introduce unobserved heterogeneity and institutions into versions for survival information. In its easiest shape, frailty is an unobserved random proportionality issue that modifies the threat functionality of anyone or a gaggle of comparable contributors. Frailty versions in Survival research offers a entire review of the basic methods within the sector of frailty types.

Adaptive Design Theory and Implementation Using SAS and R, Second Edition

Wake up to hurry on many sorts of Adaptive Designs because the e-book of the 1st variation, there were notable advances within the method and alertness of adaptive trials. Incorporating lots of those new advancements, Adaptive layout concept and Implementation utilizing SAS and R, moment version bargains an in depth framework to appreciate using a number of adaptive layout equipment in scientific trials.

Additional info for Adaptive Regression for Modeling Nonlinear Relationships

Sample text

8 simulated x values Fig. 5 1, xsim1 1, xsim2 1, xsim3 LCV: likelihood cross-validation, xsim: simulated x value The predictor 1 corresponds to including an intercept in the model a 25, slope 1, and power À7. 05, close to the true value of 5. 05, also all close to the true values. 08 % compared to the adaptively generated model, and so the true model is a competitive alternative for these data. Also, the adaptive monotonic model is the same as the unrestricted adaptive model. These results support the validity of the adaptive model selection process.

As formalized in Sect. 1, PRESS is the sum of squared deleted residuals, each computed from the observed data values for an observation along with parameter values estimated using data values for all the other observations. Smaller PRESS scores indicate better models. 4). Also, searching over the many alternative power transformations can be challenging. For those reasons, a SAS macro called genreg (for general regression) has been implemented for conducting such searches based on LCV. This macro uses heuristic (that is, rule-based) search techniques to adaptively identify appropriate power transforms of available predictors for outcomes (see Sect.

Formulation sections are not needed to understand analysis sections. 2 The Death Rate Data A data set on death rates per 100,000 for 60 metropolitan statistical areas in the US is available on the Internet (see Supplementary Materials). These data were analyzed by McDonald and Schwing (1973) and were published by McDonald and Ayers (1978) and by Gunst and Mason (1980, pp. 368–371). They are reanalyzed here to demonstrate how to conduct regression analyses that account for nonlinearity in predictor variables.

Download PDF sample

Rated 4.01 of 5 – based on 3 votes