Author Information

Karan P. Singh

Is the first presenter a student?

No

Type of Submission

Paper/Presentation

Abstract

In this paper we propose three families of functional models for analysis of dose-response data. The first family is for modeling data which have a steep sloping line for an ascending portion of the response curve and a plateau representing maximum response or a sloping line representing little response at higher application levels. The second family is for modeling data which represent a steep sloping line on the ascending portion of the response curve and a declining curvature for declining response at higher application levels. The third family is for fitting data which show an initial plateau followed by increased or decreased response, and finally a plateau representing maximum response. One of the advantages of using these families for modeling the dose-response data is that the join points of the response curves are not considered parameters to be estimated, nor their estimates considered random variables in the estimation process. The uses of the families are illustrated using them in fitting the fertilizer response data of single nutrient experiments. The problem of modeling the multinutrient response data is addressed and recently developed methods are briefly discussed.

Keywords

Families of functional models, Linear regression; Modified quadratic models; Linear-plateau models; Threshold; Plateau-linearplateau models; Bias in optimal rates; Least squares; Isotonic regression.

Creative Commons License


This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.

Share

COinS
 
Apr 30th, 10:00 AM

STATISTICAL MODELS FOR ANALYSIS OF DOSE-RESPONSE DATA

In this paper we propose three families of functional models for analysis of dose-response data. The first family is for modeling data which have a steep sloping line for an ascending portion of the response curve and a plateau representing maximum response or a sloping line representing little response at higher application levels. The second family is for modeling data which represent a steep sloping line on the ascending portion of the response curve and a declining curvature for declining response at higher application levels. The third family is for fitting data which show an initial plateau followed by increased or decreased response, and finally a plateau representing maximum response. One of the advantages of using these families for modeling the dose-response data is that the join points of the response curves are not considered parameters to be estimated, nor their estimates considered random variables in the estimation process. The uses of the families are illustrated using them in fitting the fertilizer response data of single nutrient experiments. The problem of modeling the multinutrient response data is addressed and recently developed methods are briefly discussed.